Houston Coronavirus Tracker Update 6/27/20

  • You may have noticed the "Houston" added to the Coronavirus Tracker Update title. We are expanding into Austin and Dallas. If you're interested in receiving these editions as well, please let me know.
THE COWS ARE GONE. CLOSE THE BARN DOORS. This week’s issue has probably been the hardest to write since I started the newsletter. Long time readers will know I have been very skeptical of the measures we have used to try to contain the virus. Still, while I am not surprised that we have experienced a rapid surge in cases, I am surprised to see us breach ICU capacity in the hospitals. Criticism aside, this is a very serious problem. I did not anticipate it would get this bad, I also did not anticipate:
  • The mayor and police chief leading the largest protest in the city's history with over 60,000 in attendance in tight spaces
  • That state and local contact tracing efforts would be so inept.
Yesterday we closed the bars and certainly there will be pressure for additional closings; given what we know about how the last closing effort turned out, we should expect authorities to explain how renewed closing efforts will be different to avoid this result again.

The charts below show the capacity and testing dashboards from the TMC. But first I have a chart that shows the cases/1 million population for selected states. The first 5 are states regarded as "declining cases." The last three, Texas, Florida and Georgia, are locations with "increasing cases." A couple of things here:
  • New York/New Jersey both have about 20,000 cases per 1 million population - Texas is at about 4,500. Not to minimize the wave we are experiencing or the fact that at exponential growth rates Texas could reach that level quickly. But we haven't yet and are a long way from it now.
  • Declining states have per capita infections 3-4x increasing states. Is there a baseline level at which point infections start slowing down? These states all experienced large surges in cases early. The increasing states are having that surge now after extended periods of low spread. The answers to those questions are unclear, but at the very least these charts suggest we are likely to see extended infection spread for some time.
THE MODEL'S REVENGE. I know also I said I wouldn't talk about the models again, but they made an appearance this week indirectly -- in the case numbers we are seeing. Most of the models used to justify the shutdowns back in March and April assumed an arbitrary cut-off that pushed the majority of cases and fatalities outside the model horizon. A couple of professors at Carnegie-Mellon reverse engineered one of the models and found a two-month shutdown nationally produced several million more cases than actually doing nothing.

The chart below shows one version of a model that was published in the New York Times and claimed that two months of social distancing reduced cases from 128 million to 14 million. The paper describes several other models in publications like the Lancet that reach similar conclusions. It would be hard not to make this decision in light of those facts.
The professors were able to hack the model and run its scenario through October. What they found was that this same model projected 194 million infections, 50% more infections and 50% more fatalities than the non-lockdown scenario simply by extending the model's time horizon. They saw similar results in multiple other models.

I wonder if this is not in part what we are seeing now. This is yet another reason to be wary of pandemic modeling.
I DON'T KNOW HOW BIG THIS THING IS GOING TO GET. I have shared many stories indicating the coronavirus infection has spread much more widely than we have been told; most notably through extensive sewage viral loading studies conducted all over the US and Europe. This week the CDC has announced COVID-19 cases in U.S. may be 10 times higher than reported, in conjunction with a study from Penn State that reveals we may have had almost 9 million infections in the US in March, before the shutdowns. Why is this important? The CDC's own Pandemic Response Plan says

“the effectiveness of pandemic mitigation strategies will erode rapidly as the cumulative illness rate prior to implementation climbs above 1 percent of the population in an affected area.”

If the 8.7 million number is accurate, that's almost 3% of the US population - and the percentages were probably higher in large metro areas. That's nearly 3x above the ceiling of infection the CDC's own plan said was required to be effective.

The news is not all bad: A recent study shows that less than half a population needs COVID-19 infection for herd immunity. If true, immunity can be achieved much more quickly than originally thought. It's not an impossible concept - in 1957-58 the influenza pandemic burned out after infecting only about 25% of the population.

Another is the declining death rate - which is largely a function of who's getting sick. In Texas the infection numbers are now very high in the 20-39 age cohorts which may mean less serious infections and fortunately fewer fatalities.

The final positive note I found out of South Florida. When elective procedures resume, COVID isolation wards were filling up with asymptomatic patients who were there for other procedures. For example, a patient with appendicitis who tested positive but was not otherwise sick, would recover in a COVID ward rather than a surgery ward. I've heard anecdotally of similar cases here although I can't quantify them. But the age of new infections and rise in share of COVID patients at least suggest the possibility.
RESTAURANT WATCH. Chase Bank says "card-present" restaurant spending is highly predictive of infection spread, much like open table data. Correlation is not causation of course, but definitely bears watching.
Durable goods are starting to recover, although still below 2019 levels.
New unemployment claims have flattened out, but continuing claims remain sticky with little movement back to employment. And the rise of infections imperils employment gains in Texas, Florida, Georgia, etc.
EXPECT UNEMPLOYMENT REVERSALS. Unemployment claims are still unacceptably high - as of June 13, they were about 22,000 in the Houston MSA. The same week last year was just under 4,000 claims. They have gone down, however, from almost 80,000 at its peak. Bars, restaurants and hotels employed about 330,000 people in the Houston area before the pandemic. With the surge and associated closures, I would expect to see these declines reverse as the trend describes above.
HOME SALES IN FULL REBOUND. Although the market was greatly affected during during April and May when the pandemic was at its peak, sales have increased sufficiently in the last month that June is up YOY. All data from HAR MLS through June 25.
Large declines are visible on a weekly basis, particularly weeks 13 (3/23) and 17 (4/20), homebuyer interest in 2020 as expressed through sales has been fairly consistent compared to 2019.
Averaging pricing for new homes has completely recovered from the pandemic decline and is now ahead of last year.
Pending new home sales in MLS remain strong - and are increasing by almost 100 contracts per week since June 1.
MATH IS HARD. Congratulations to Bill Frazier, Bob Douglas, Art Nevid, Jim Jenkins, Michael Bloom, and Mike Miller for their correct guesses. Gift cards are in the mail. Right triangles are defined by the Pythagorean Theorem, which holds that the lengths of each side are represented by a 2 + b 2 = c 2. In the sign to the left, the longest side, c, is 6 feet - which makes the other sides 4.24". So what's wrong with the sign advising you to stand 6' apart? It shows four individuals standing 4' 3" from each other.
THERE'S A REASON IT'S CALLED THE WATER OF LIFE, AFTER ALL. Whiskey distillers struggle to recover from COVID-19.
Are you looking at new deals? Want to understand what's going on in the market today? We're here to help, even if it's just a short conversation.

Stay well.

Scott Davis
Location Strategy, LLC