COVID-19 presents a technical problem we don’t often encounter. Almost all of my (career) work has involved a fixed data set run through a standard (customizable) model to produce a deterministic (financial) result. Here we are trying to get a handle on a pandemic developing in real time during our analysis, and our data sources are (1) news reports and medical theories, many of which are in conflict, and (2) US and global data which is apparently subject to all sorts of manipulation and suppression (Brazil announced today that it will no longer share its pandemic statistics publicly). In this situation more than ever, context is important; for example, given the societal differences, I would predict/guess that Brazil’s actual data (if we ever see it) will be a better measure of the toll involved in herd immunity than Sweden’s will be.
Some details on the different types of COVID-19 testing and how they work: https://www.usatoday.com/story/news/health/2020/06/06/coronavirus-test-covid-19-tests-increase-plan-lacking/3138440001/ (link replaced 6/8/21). I’ve also read that the Oregon State Penitentiary has one of the largest outbreaks in our state (123 cases) but only 4% of the facility’s population has been tested. This just seems stupid – easiest place in the state to contact trace and isolate.
Some surprising medical theories on increased male mortality with COVID-19: https://bestlifeonline.com/coronavirus-baldness/ (link replaced 6/8/21). And check out the links in this article; turns out certain male physical characteristics reflect particular male hormone levels, which are in turn implicated in the severity of the illness the virus causes based on its ability to enter your cells.
The unemployment rate’s miscalculation error: https://www.cnbc.com/2020/06/05/heres-why-the-real-unemployment-rate-may-be-higher-than-reported.html . The April rate should have been about 20%; the May rate at least 16.3%. More broadly, it appears many statistical models out there have problems when the inputs don’t move incrementally … as Black Swan events become more common, more modeling shortcomings will undoubtedly appear.
WSJ: Banks move more online: https://www.wsj.com/articles/people-arent-visiting-branches-banks-are-wondering-how-many-they-actually-need-11591531200 . This means accelerated staff reductions at banks and in other businesses as well. At my (chain) grocery, people increasingly bag their own groceries at the automated check-out rather than go through the cashiers to prevent others from handling their food.
WSJ: But can we automate contact tracing?: https://www.wsj.com/articles/why-google-and-apple-stores-had-a-covid-19-app-with-ads-11591365499 . Shouldn’t we be regulating internet providers as utilities?
Following up on the debate within the New York Times editorial department: The editor has resigned: https://www.theguardian.com/media/2020/jun/07/new-york-times-editor-resigns-tom-cotton-oped-protests (link replaced 7/28).
An upbeat note: Recent primary voting expansions “largely nonpartisan” (meaning bipartisan): https://thehill.com/homenews/campaign/501384-turnout-surges-after-states-expand-mail-in-voting (link replaced 6/8/21) (and I wrote the intro comment before receiving Jeff Petertil’s comment).