Coronavirus Tracker Update 5/9/20
Previous editions: May 2, April 25, April 18, April 11, April 5

TURN ON, TUNE IN, DROP OUT. What major 20th century cultural event took place during a pandemic? Answer is below.
"THEY KNOW THEIR CODE CAN'T WITHSTAND SCRUTINY." The fight now is for the economy, so the rest of the coronavirus business is below. But I wanted to highlight a former Google software engineer's review of the code from the Imperial (UK) College's model that predicted 2.2 million deaths in the US. It's devastating, and you don't have to understand programming to understand it. But if you don't want to read it, read this:

All papers based on this code should be retracted immediately ... I’d go further and suggest that all academic epidemiology be defunded. This sort of work is best done by the insurance sector. Insurers employ modellers and data scientists, but also employ managers whose job is to decide whether a model is accurate enough for real world usage and professional software engineers to ensure model software is properly tested, understandable and so on. Academic efforts don’t have these people, and the results speak for themselves.
HOW DOES LOCATION STRATEGY HANDLE MODELING? We have moved all of our forecast models to a code platform where we can show you the specific code that generates each model component without having to trace column after column of obtuse Excel formulas. If we can use one of these transparent models to help you on a project, we do feasibility studies, MUD studies, project and product positioning, asset location and market opportunity surveys, please let me know.
IT'S THE ECONOMY, STUPID. You may have seen Dr. Gilmer's webcast where he announced a potential loss of 83,000 jobs for Houston in 2020. I've just seen it myself and will look into this for next week. But even Dr. Gilmer acknowledges "risk" to that forecast. I think we have to focus on several questions to adequately anticipate how high job losses will get, and how they will affect our industry:
  1. How many job "losses" are really furloughs, with the anticipation that they will brought back? The labor department graph below suggests 18 million (90%) are on furloughs - that's still 9.0% unemployment nationally, but not Great Depression-levels of unemployment either -- much like the impact analysis I showed last week.
  2. Health care industries are #2 in job losses in Texas. How quickly can this industry recover from the "protection" we've been giving it?
  3. Government decrees don't decide when businesses will open, business owners do. How soon will that happen? Mark Cuban recently did a survey in Dallas of retailers and restaurants and found only 36% were opening for "reopening weekend."
  4. How many of the permanently lost jobs are in categories like accomodations and retail trade, that are less likely to affect homebuilding?
  5. Last but not least, what will it take to get a large sector of the public with real and imagined fears of the coronavirus to return to regular commerce? That same survey uncovered serious issues the business community will likely have to address.
  • ARE BUYERS DELAYING CLOSINGS? There are still over 2,700 pending new home sales in the HAR MLS. We wanted to take a look and see if buyers were extending closing times on new contracts compared to what's sold this year. For 2020 sales, the average contract period was 79 days; for currently pending transactions, it's only been 47. But as so many other things in our market, there is a sharp division in the $400Ks. The blue line below shows the average contract period for closings by price range. The orange period shows the time between the pending date and May 8. Above the $400Ks, properties on average have been under contract longer than similarly priced properties that have already closed this year. Some of them a lot longer. The good news is this is a smaller segment of the market, and by this metric one of our most important segments, $200-299K looks to be in excellent shape.
  • Note the trend: the second graph in this series shows contracts written (sales + pending) for March-April-May for 2019 and 2020. I have not included terminations from either period because I've found the pending dates unreliable. While they don't cover every transaction, the data suggest that new contracts were off about 15% in April. Also, note that in May (through May 7), we've already written one-fourth as many contracts as in the entire month of May 2019.
Source: HAR
  • SPREAD BETWEEN SALES AND CANCELLATIONS WIDENS: I noted last week that April sales would end up down 20% over last year, which they did coming in at 19% below 2019 levels. For YTD we are still ahead of last year but won't end up that way. But if we continue to see normalization in the economy, we should see it in the housing markets as well. Since week 17, we've seen a broadening in the spread between sales and cancellation rates, and the spike in cancellations in weeks 14-16 compared to 2019 seems to have abated.
  • AROUND THE STATE: I took a few minutes to check in around the state to see how Houston compared to other metros based on new sales, existing sales and new listings. Houston appears to be in about the middle of Austin and Dallas in these metrics. Honestly I'm not sure why we appear to be outperforming Austin - the Houston job situation is much worse. Source: HAR, NTREIS, ABOR.
TMC showed a small increase patients in the ICU with COVID-19. That leaves 560 beds capacity, and 1,200 beds surge capacity.
IT'S UP TO YOU, NEW YORK. You may have recently seen this chart which showed cases falling in New York and rising everywhere else in the country. And the chart is true. But when you look at the same data on a per capita basis, it looks exactly like what the inset graph from the CDC (with some annotations) shows: New York failed to manage the spread of its infection in a way the rest of the country did not. More explanation here.
THE SHUTDOWN CEILING? The graph below from the CDC illustrates the trend from above. The CDC is no longer releasing the "onset date" series we were tracking, so we've had to rely on the "reported date" series. The graph below shows daily new cases - and they appear to have followed an upward sloping curve before topping out on April 7 and since then have essentially remained flat. It is not a function of limits of testing; testing capability tripled during the flat period.

This is exactly what we should have expected with most shutdowns starting around March 20 - and an incubation period of 14 days or less in 99% of cases. While there is some argument as to whether this represents a success or not, the trend suggests that by mid-April we may have hit a "ceiling" on the effectiveness of shutdown strategies, at least nationally.
CORRELATION IS NOT CAUSATION - IS THE OPPOSITE TRUE? Yinon Weiss is a technology entrepreneur, Harvard MBA and an Iraq war veteran. Not exactly the qualifications you see in epidemiology experts. But he's written an amazing series of posts on Medium that caught my eye.

The chart below shows the fatality rate (deaths/MM) after 21 days from SAH orders vs how quickly those orders were adopted relative to the state hitting 1 death/MM. It's a small chart; the R-squared is 0.0119. Why 21 days? 99% of cases develop within 14 days of exposure. There is no general correlation between how fast a State shut down and how many people died in the first 3 weeks following an early mortality milestone. Weiss has some other charts and concludes:

We need to better understand the cost/benefit of different kinds of social distancing. This is important as our nation begins a marathon of mitigation against coronavirus, and in this struggle we need to better understand the benefits and the trade offs of all the measures we can take. Allowing the purchase of jewelry and lottery tickets but not gardening supplies is an example of how misguided some of our government decisions have been. Next time somebody seems to panic and wants a shut down, we should remember that there is likely a long list of less costly measures that may be even more effective before we jump to the most drastic extreme.
DO AS I SAY, NOT AS I DO. I noted Imperial College earlier; you may remember Neil Ferguson as the director of the university's forecast team. Now you can also remember Dr. Ferguson as the epidemiologist who had to resign his position as a government advisor because he violated lockdown rules by having his lover visit his home multiple times. There was no word at press time if Dr. Ferguson plans to update his forecast to include the impact of requiring epidemiologists to comply with their own recommendations.
NOT MUCH SOCIAL DISTANCING AT YASGUR'S FARM. Woodstock was held in August 1969 - and planned in late 1968 and early 1969, during the Hong Kong influenza pandemic. Reportedly 400,000 people attended. There were also more than 500,000 US troops overseas in Vietnam. “In 1968, the H3N2 pandemic killed more individuals in the U.S. than the combined total number of American fatalities during both the Vietnam and Korean Wars.” The pandemic took over 100,000 lives when the nation's population was only about 200 million - the equivalent of about 165,000 today - and yet is hardly remembered. It's not like the US was unfamiliar with disease; the Asian Flu Pandemic of 1957 had killed the equivalent of 200,000 today just ten years before - and 1918 would have been within living memory at the time. The CDC was founded to eliminate malaria in 1946, which had US death rates of 40/million pop throughout the 1930s and was only eliminated in 1951. And the polio vaccine in 1955 ended outbreaks of the disease which crippled 35,000 children per year.

Again, still doing webinars. The webinar covers the national and local economic situation and Location Strategy's current forecast for Houston, and I find I've been running about two weeks ahead of most media reports and government actions.

Stay well.

Scott Davis
Location Strategy, LLC