Sunday, May 24, 2020

Coronavirus dashboard: emphasis on testing


 - by New Deal democrat

I want to focus this edition on testing issues.

While the seven day average number of deaths continues to decline:


The seven day average number of new infections has leveled off:


The average number of daily tests *may* also be leveling off again in the past few days:


What is worse is that the number of new cases in the US has only declined -22% from its peak in the last 45 days. Meanwhile, even hard hit countries in Europe like Spain have seen a -90% decline from peak in new cases over a similar timeframe:


Even hardest-hit Belgium has had a -80% decline from peak (not shown).

The pessimistic explanation for this is that the US’s response, on the whole, has always been half-hearted at best. That may be the case, and at very least it is part of the explanation.

The more optimistic explanation - and I think this is the main part of the answer - comes back to a poor level of testing. 

As I pointed out yesterday, in the matter of testing the US started late, with an abominably low number of tests in March and early April, and only in the past week to 10 days equaled and finally slightly surpassed the level of testing per capita in comparable countries in Western Europe:


On a cumulative basis, the US come in dead last for per capita number of tests administered (We’re number 178!!!)(column at far right): 


Altogether the US had administered 38,000 tests per million people, while Italy has tested 51,000, Belgium 61,000, and Spain 65,000. 

A study of 366 US counties just published by the Surgo Foundation indicates that the US has probably missed 40,000 deaths or more, especially in rural counties, by its failure to diagnose infections:

Our data suggest that we may be especially undercounting COVID-19 deaths in Southern parts of the US. In more than 200 of the counties we examined, excess death rates were between two times to 30 times higher than reported COVID-19 death rates.
On average, excess death rates are 30% higher than reported COVID-19 death rates.

Here’s a graph from the study showing the trend:


In general, the higher the rate of coronavirus infection, the higher percentage of deaths that have been missed. Further, note the above is deaths only, not even total diagnoses. Probably the number of diagnoses missed has been about 5x or more the number of deaths missed.

In short, in March and April the US was missing a huge share of coronavirus infections, that have been increasingly picked up since testing increased beginning in late April.

If you go back and take a look at the fourth graph in this post above (comparing US cases per capita with Spain), you can see that there was a similar leveling off in the number of new cases during the middle portion of April just after a big increase in testing. I suspect that the current leveling off in new cases primarily reflects that the US as a whole finally ramped up testing enough to catch these cases.

The silver lining is that about half of the US is probably near, if not at, the point where it is performing enough tests to begin transitioning to a “test, trace, and isolate” regimen. Here’s a graph from The Johns Hopkins University’s Coronavirus Resource Center,  making use of the WHO‘s recommended positivity rate of 5% or below for 14 days before implementing a “test, trace, and isolation” regimen, indicating that 26 US States and territories have reached that  point in testing. The other 26 have not:


The huge spike at the right end of the graph is Puerto Rico, which is an order of magnitude off the number of tests that need to be performed, followed by Maryland. 

It will probably take about another week to see if the trend in new cases starts to decrease again, or whether the reckless “reopening” of many States begins to show up in a second wave of new infections.

Saturday, May 23, 2020

Weekly Indicators for May 18 - 22 at Seeking Alpha


 - by New Deal democrat

My Weekly Indicators post is up at Seeking Alpha.

Many of the indicators are bouncing off their worst levels, especially those like mortgages that are affected by lower interest rates. On the other hand, employment losses continue to spread out.

As usual, clicking over and reading not only should be educational for you, but rewards me a little bit for my efforts.

Friday, May 22, 2020

Bad news and good news on coronavirus; plus, implications for Election Day


 - by New Deal democrat

No economic news today as we head into the Memorial Day weekend, but there are a few coronavirus and economic/political developments of note.

First, the bad news: the declining trend in new diagnosed cases of coronavirus in the US has stopped in the past week. Instead new cases have leveled off. Here’s a graph from Conor Kelly’s excellent tableau coronavirus dashboard page:



Cases in the US outside of NY actually increased slightly (2%) in the past week.

Now, the good news: as also shown in the above dashboard, testing has continued to increase dramatically, up 18% again in the past week.

The jury is still out on whether the slight increase in new cases in the past 4 or 5 days is because some States recklessly “reopened” without justification, or whether it is just an artifact of increased testing. There was a similar small increase in new cases right after testing started to ramp up around April 20, but it lasted less than a week. It will take about another week to know which explanation for the end of the decline in new cases is the better one.

Even better good news about testing: the US is no longer lagging compared to other major developed countries. The below graph shows the 7 day average in US testing per capita compared with those countries in Western Europe that also had major outbreaks, plus South Korea:



As you can see, South Koreas started large-scale testing immediately and was rewarded with “crushing the curve,” allowing the country to go back to nearly normal activity. The US did awful even compared with Europe until about April 20, when testing in the US ramped up significantly. In the past week, testing in the US finally surpassed all of the major countries in Western Europe except for Belgium (the worst hit on a per capita basis).

Of note, at their worst on a per capita basis both Belgium and Spain had worse outbreaks  than the US does now:



This strongly suggests that, with proper staffing, much of the US could begin an aggressive tracing program to help tamp down new infections more.

In that regard, there is a new site called “Test and Trace” which is tracking exactly that. Here’s their map, as of May 20, of which US States are able to begin a successful test, trace, and isolate regimen:



My one criticism is that the site measured “planned hires” for tracers in each State, as opposed to *actual* hires.

I continue to think that it will be crucial for at least one medium sized or major State to “crush the curve” to show other States that the plan can be successful - particularly if new outbreaks begin in States that have recklessly relaxed restrictions.

Finally, on a related political/economic note, the April index of Leading Indicators was reported yesterday, and as expected it was another bad month, down -4.4%:



Since the LEI tells us where the economy is likely to be in the next 6 to 8 months, this tells us that conditions are still going to be poor on Election Day. And every econometric election model says that means the incumbent will be defeated.

Thursday, May 21, 2020

Initial jobless claims: employment damage continues to spread


 - by New Deal democrat

Now that there is more than one month of data from initial and continuing jobless claims since the coronavirus lockdowns started, we can begin to trace whether the economic impacts of the virus are being contained, or are continuing to spread out into further damage.

Nine weeks in, it appears that, insofar as employment is concerned, the damage is continuing to spread.

First, let’s look at initial jobless claims both seasonally adjusted (blue) and non- seasonally adjusted (red). The non-seasonally adjusted number is of added importance since seasonal adjustments should not have more than a trivial effect on the huge real numbers: 


There were 2.174 million new claims, which after the seasonal adjustment became 2.438 million. This is a slight decline from last week’s number which was revised down to 2.687 million. 

By now, virtually all of the people laid off due to the initial lockdowns in March and early April should have already applied for benefits. Further, last week was the second week after some States “reopened.” Thus these new claims are almost certainly primarily represent the spreading second-order impacts of the coronavirus shutdowns. In other words, this is evidence that new economic damage have  continued to spread, and in a very large way.

Next, looking at continuing claims, which lag one week behind, both the non-seasonally adjusted number (red), and the less important seasonally adjusted number (blue) rose:


This tells us that, as of two weeks ago, there were not enough callbacks to work to offset the spreading new damage. If “reopening” leads to a significant new upturn in cases - something that may have begun in the past week - this will only get worse.

Bottom line: confining my comments strictly to the economy, while there have been significant or small rebounds in many of the series, the news on employment is not just bad, but it is still getting worse, albeit getting worse at a slower rate.

Wednesday, May 20, 2020

Housing permits and starts plummet in April, but mortgage applications suggest big rebound in May


 - by New Deal democrat

At some point it is going to be safe for the economy to be completely reopened. When that point comes it would be nice if the leading sectors of the economy have already been priming the pump for a consumer rebound.  As usual housing is the most important long leading sector in that analysis.

As expected, housing tanked in April. But it is likely setting the baseline for improvement in the coming months, as new record low mortgage rates have brought out new buyers, as shown by new mortgage applications which as of this morning are only -1.5% below where they were last year at this time.

This post is up at Seeking Alpha. As usual, clicking over and reading should be educational for you, and helps reward me a little bit for my efforts.

New coronavirus cases vs. testing in “reopened” States


 - by New Deal democrat

Are new coronavirus infections increasing in States that “reopened” on or about May 1? The jury is still out. The number of infections is up in 4 of the 5 biggest States that have done so, but so are the number of tests. The likelihood that most or all of the increase is an artificial of an increase in testing depends on the date on which you start your comparison.

I haven’t been able to find graphs that nicely show both tests and positives together, so let me just show you 2 separate graphs of the 7 day average in number of new cases diagnosed vs. the 7 day average in testing for Texas, which is the state with the biggest number of new cases.

First, here are new coronavirus infections, which are up 42% from April 30:



But here are new tests performed, which are up 120%:



To save space, I won’t show them, but the trends are similar in the 4 other States, except in Georgia, where there has been an actual decline in new cases reported since April 30. 

Instead I’ve prepared the below chart of both the increase in tests and increase in cases for the 5 target States. Additionally, as you can see in the graph of Texas above, there seems to have been a renewed pulse upward in new cases since May 10. Since there is a 5 to 10 day lag between exposure and the onset of symptoms, this may be what is showing up. A similar pattern seems to be occurring (but could be just noise!) in several of the other States, so the below graph also compares test increases to new infections dated from May 10 as well:

State Test increase
From 4/30
Case increase
From 4/30
Test increase
From 5/10
Case increase
From 5/10
AZ157%51%-41%15%
FL35%12%4%21%
GA135%-2%65%-13%
TN41%-1%12%36%
TX120%42%77%19%











All 5 States show bigger increases in tests than in new infections when dated from April 30. Dated from May 10, 3 of the 5 States - AZ, FL, and TN - show a bigger increase in new cases than in testing. Georgia remains a complete outlier, but has been shown to have important data integrity issues.

It is certainly possible - indeed likely - that many people, especially seniors, are probably continuing to practice safe distancing and mask wearing habits, despite the local governments’ lifting of restrictions, and this is probably at least a partial explanation of the data. But the virus is impervious to ideology. To the extent people in the “reopened” States are engaging in unsafe behavior, we should expect new infections to increase. Nevertheless it is safe to say that nothing dramatic has happened in the first three weeks of May.