Today is a newsy day. Here is the crux of all of today’s update: it’s going to be worse than you think, but that isn’t as bad as you think, as long as we keep doing the right things.

First, a core assumption in our own projections: we assume that COVID-19 is bi-partisan, non-racist and doesn’t respect boundaries or political beliefs. We believe that currently there are not 10 or 5 or 3 different ways to react to protect people from COVID-19, there is just one. As a result, when we see patterns in how policymakers respond in different jurisdictions we assume that typically others will soon follow because that is the way to respond.

We’ve seen this play out the in closing of schools so far, stay-at-home orders, travel restrictions, and other forms of social distancing. This assumption will be relevant later in today’s update.

Second, let’s take a look at the big picture from two different angles before we get into the weeds. This may be the most important part of today’s post.

How bad is it?

Let me just quote a Facebook friend in Tennessee here:

Presently in the U.S. you’re more likely to die of Coronavirus in a given day than all but 2 of the Top 10 Causes Of Death.

Let me explain…

In the past 24 hours [March 31-April 1] we had somewhere around 900-1000 COVID-19 deaths.

Here are the CDCs numbers for the Top 10 Causes Of Death, with total number of deaths per year and then broken down to per day.
1. HEART DISEASE 647k –> 1772 per day
2. CANCER 609k –> 1670 per day
3. ACCIDENTS 169k –> 463 per day
4. COPD 160k –> 438 per day
5. STROKE 146k –> 400 per day
6. ALZHEIMER’S 121k –> 331 per day
7. DIABETES 83k –> 227 per day
8. FLU 24-62k –> 100-258 per day
9. KIDNEY DZ 50k –> 136 per day
10. SUICIDES 47k –> 128 per day

When it gets to 2200+ per day {and it will…soon} you’ll be more likely to die of coronavirus in a given day than all but the Top 2 Causes combined.

Buffy Jay Cook, April 1, 2020

Of course, we live with all ten of these things and don’t call it a national emergency, but none of the other ten double their deaths every 2-3 days. When we had 8 suicides in one day in Knox County our county mayor and congressman both took notice, as they should have. The same urgency applies to COVID-19.

Axios asked today if you can remember what March 1 looked like, when there were only about 7,000 cases — total — outside of China. The US alone added 30,000 cases today.

And it’s supposed to get worse before it gets better. At 30,000 new cases, today is probably the best day we’ll have until mid-May.

Red line is constant 1,000 deaths/day

You’re probably thinking we’re trying to scare the bejeezus out of you. We’re not, this is just what the data and models predict, right now.

How bad will it be?

Our director of operations noted today on a call that the total number of US deaths the currently projected high end of reasonable estimates roughly matches the number of people killed in an hour in 2004 when a tsunami hit the Indian Ocean.

In 2017 the US had roughly 2,800,000 deaths according to the CDC. We’re on track to have a year that comes in 3.5% to 10% above that in total deaths [1]. We don’t want to minimize any given death — each and every life is precious. Yet, in the grand scheme of things this isn’t the zombie apocalypse.

…but

Remember that these current projections are based on all the shutdowns and social distancing we are currently practicing. Change our behavior, change the results. Right now, our behavior is changed for more favorable results. Go back the other direction and we’re in a spot where that peak — two to three weeks out — is much higher and much farther away.

We are in a situation where cases/deaths are doubling every few days. The lower the number we stop the doubling at the less bad it will get. By a factor of 2 or more!

In sum, it’s going to be worse than you think, but that isn’t as bad as you think, as long as we keep doing the right things.

Important News Items

Georgia called off school for the rest of the year today, as did Indianapolis. You recall our assumption at the top? We expect to see a lot of other places follow suit. We’re all for that, it makes more sense to acknowledge reality now and plan for it than to keep introducing new delays every two weeks.

In Tennessee Governor Bill Lee issued Executive Order 23 which changed Executive Order 22 from “urging” people to stay at home to “requiring” people to stay at home, unless engaged in essential business. As we noted when Executive Order 22 was signed, the list of “essential” businesses is very large. We expect that this will close out the remaining recreation (golf courses, parks, etc.), we’re not sure what else.

News reports cited the same Unacast social-distancing scorecard we referenced a few days ago as playing a part in Gov Lee’s decision. In short, Tennesseans aren’t staying home. Honestly, just getting out the last few days (by car) has produced enough data to make that clear to us.

Mayor Jacobs made it know that he wasn’t really happy about the order. Here we will simply note that there is a long line of policymakers ahead of him that were against a stay-at-home, until they weren’t, including: President Trump, Gov. Lee, UK Prime Minister Boris Johnston, Gov. Cuomo of NY, Gov. Stitt of OK, Gov Scott of FL. We’re sure we don’t have them all. The point? The data seem to have a way of convincing principled policymakers that these orders are entirely justified after things get to a certain level.

Dr. Fauci thinks that all states should have a stay-at-home order right now.

There are a number of articles out today that indicate that banks are not ready for the flood of SBA 7A PPP applications that they are set to start receiving tomorrow. There are 30,000,000 small businesses in the US. In 2019 the SBA issued 47,000 7A loans. We are 100% confident that no bank or government agency is ready to receive a 5000% increase in a single day compared to all of last year [2]. And while everyone is trying to figure out working from home.

At JM Addington we’re working on ensuring that we have other sources of cash (e.g., a line of credit) available while the loan works its way through and are working on preserving all the cash that we can in the interim. Your author spends close to a quarter of his day every day on this currently. Also, the guidance on supporting documents changed since it was first issued. See First Horizon’s info here.

Self-employed persons can now apply for Pandemic Unemployment Insurance in Tennessee. This is huge. If you know a small business owner, “gig worker,” or anyone that doesn’t work for a company who is out of a job send them to jobs4tn.gov and the FAQ here.

Thanks to the West Knoxville Farragut Chamber for this tidbit.

Also, we’re really proud of our state for stepping up in this way. It’s the right thing to do.

I am self-employed, when can I start applying for Pandemic Unemployment Assistance?

You should apply for unemployment insurance now, even if it the system shows no earnings. Once the federal government provides states with guidance on how to administer the Pandemic Unemployment Assistance program, Tennessee will start providing those benefits.

https://www.tn.gov/workforce/covid-19/faq.html

A DENSO employee has tested positive for COVID-19. The person was last at the facility March 25th, “We are taking this matter very seriously and have closed Building 102 and other areas in which this person may have been, so a professional third-party cleaning service can thoroughly disinfect the building,” a spokesperson said. Given the contagious nature of COVID-19 we are not optimistic that these measures will be adequate for DENSO.

Memphis is reporting that the first wave of COVID-19 patients is hitting the hospitals there. “So that volume is already coming into our hospitals in a big way. And the public just doesn’t know that, I believe.” Again, watch how policymakers respond in TN when media coverage in Memphis looks like New York or Italy. (Also, remember, that by volume it’s not comparable, but expect some compelling photos and video that would make it seem so.)

Numbers

We’re not even going to update today. New cases and new deaths are up, unemployment numbers are up, etc., etc. None of that is news.

We think the more useful measure for today is: how far into this are we? This graph gives us the best answer:

Note that Tennessee is set to peak about a week later. We haven’t seen sub-state regional breakdowns.

International Numbers

Italy’s new cases are still, overall, going in the right direction.

Source: Worldometers

Spain, Germany, France and the UK are all still on the upward curve, or at least not clearly trending down.

Hong Kong shut down restaurants and pubs for two weeks. Hong Kong’s population density is similar to LA, they are obviously closer to COVID-19’s outbreak than the US, has the population as Tennessee and only 1/3 of the cases. And only now had to close down the restaurants and pubs. So what? There are clearly more effective ways for fighting pandemics than the route we’ve taken in the US. Hong Kong may be the world’s leading example.

Other

Thanks to those that have shared kind words or liked these posts. We’re doing out best to put out data-driven analysis, each one of these takes about 2 hours. It’s helpful to know that they are being read.

If you want copies of the Excel sheet and PowerBI Reports we use to put these together email us, info@jmaddington.com. Right now, our PowerBI combines data from JHU, NYT, COVID Tracking Project, IMHE and the TN Department of Health. Most sets are updated daily.

[1] Not correcting for the fact that a large number of the COVID-19 deaths will hit the same people most likely to die of other causes this year.

[2] Obviously we don’t know how many businesses will apply. This number assumes half of all small businesses apply. Even if you drop it down to 1% of all US small businesses that is over 6 times the amount of 7A loans the SBA processed last year.

Today is about map reading, and accuracy.

The biggest we face — at least at JM Addington — in projecting out what the next few weeks look like is the lack of consistent, quality data. First, a really quick primer on the stats we are watching:

New hospitalizations: Number of new people admitted to the hospital. This is not the number of beds in use as most people will stay more days or weeks. Lots of new hospitlizations indicate pressure on the entire healthcare system.

New deaths: New number of deaths attributed to COVID-19. Unlike cases deaths don’t depend on someone running a test. However, they can still be undercounted (someone dies without ever having a test) or overcounted (a person with multiple health issues happens to get COVID-19 and pass away, but they may have regardless). Best used in combination with new cases.

New cases: New, confirmed (i.e., tested) number of cases of COVID-19. Best used in combination with new deaths.

Tests: Number of tests performed. Typically, one test = one person although that doesn’t have to be true. The data on testing in the US is not consistent. Tests are often broken out into pending/positive [cases]/negative.

Ok, let’s roll.

We have two significant issues with testing, at least as far as the data tell us. First, we haven’t tested that many people.

While testing is clearly ramping up, testing less than 1/300 just isn’t where we need to be today. Just as significantly, those tests are performed in heavily concentrated areas:

New York, a state with less than 9 million people, makes up 18% of total US tests. California has 8%, and so on. On the one hand we certainly expect that NY has the most tests as they also have the most cases far and away. But if we test them the most we’ll also find more cases. It also overweights the data for the US as a whole. In some fields, this could be corrected. We haven’t seen any convincing evidence that such statistical correction (cleaning up dirty data) is yet possible with the COVID-19 tests.

We wanted to see how those IMHE projections were mapping up with reality. These are some of the projections that the White House has been relying on when giving guidance over the next month. IMHE updates the model as new data comes in so we compare current data against their first mode (March 27th) and today’s:

The blue lines are IMHE’s worst-case scenario, best case scenario and their splitting-the-difference scenario for new deaths by date. The pink is where we actually are, you can see that it is tracking closely.

Today’s updated model looks remarkably similar:

Actual New Deaths from The COVID Tracking Project

Four days isn’t a long time out for a model to predict but it’s enough to give us some confidence that it is off to the right start. The new hospitalizations (today’s model), in contrast, seem to be off by a fair bit. It’s also possible that the data is lagging.

Actual New Hospitalizations from The COVID Tracking Project

As long as these models show accuracy we’ll keep them up as a way to score where we are on the curve.

It’s also important to remember that these curves can be highly localized, that is, TN is different from NY and even Knox County is different from Davidson County. All of the US-as-a-whole charts around COVID-19 are heavily weights towards NY right now. That does not mean to dismiss the data, just remember that portion of the data is telling you a different story.

In Tennessee, we continue to have two counties dominate the number of total cases — and tests:

Candidly: the numbers for the Knox County area continue to surprise us. While still increasing exponentially, the absolute number of cases is still incredibly small, on the whole.

Only 157 positive cases as of today in the 9-county area.

It is, almost, unbelievable.

That also brings us back to our beginning problem: how good is the data? 700 tests out of 200,000 people (Knoxville) wouldn’t be an awful statistical sample if it were randomized. At the same time, the signals coming from the Knoxville County Health Department and Mayor Kincannon are, subtly, that they believe these numbers are too low as well.

Data from the New York Times.

We don’t have a good answer for you right now and we’ll refrain from guessing. We’re just going to present the data that we have.

International Numbers

Italy’s new cases are still, overall, going in the right direction.

Source: Worldometers

Spain, Germany, France and the UK are all still on the upward curve, or at least not clearly trending down.

Other

Thanks to those that have shared kind words or liked these posts. We’re doing out best to put out data-driven analysis, each one of these takes about 2 hours. It’s helpful to know that they are being read.

If you want copies of the Excel sheet and PowerBI Reports we use to put these together email us, info@jmaddington.com. Right now, our PowerBI combines data from JHU, NYT, COVID Tracking Project, IMHE and the TN Department of Health. Most sets are updated daily.