Elon Musk has said that he thinks not of building cars but of building factories that make cars, the machine that makes the machine. You can see what he is on about in this new video of the first Gigafactory.
Three points of note. The factory was up and running in 10 months. There are lots of robots, in a factory in China–that tells you a lot about Chinese wages and the productivity of robots today. Tesla is building Gigafactory Berlin and Gigafactory Texas next.
Here’s the latest Economists in the Wild video featuring Amy Finkelstein, Tamar Oostrom and Abigail Ostriker discussing some of their research (with Einav and Williams) on breast cancer screening. It’s a good video for illustrating how the tools of economics can be used to study a startling wide variety of problems.
I’ve been pounding on the need for fast, frequent testing but it’s clear from some of the comments to The Beginning of the End that I have failed to convey some fundamental points. A seemingly sophisticated objection is to note that given background prevalence rates even a fairly specific test will result in a high fraction of false positives among those who test positive. (This is the standard Bayesian doctor puzzle.) It’s nice to see people doing the Bayes calculation but some of them are then drawing the wrong conclusion. Let’s spell it out.
Suppose that the numbers are such that 50% of the people who test positive actually are negative. That sounds bad and it’s not great for a diagnostic test but it’s good enough to be a massive help in a pandemic. To see why, just imagine that you could easily identify people who had a 50% chance of being infectious. That’s a very useful piece of information! If just this group were to intensify social distancing for a week or two the pandemic would end quickly.
In essence, testing allows us to target and intensify social distancing on the people who are most likely to be infectious. Suppose that 10 in 1000 people are infectious (a 1% infectious rate) and that all 1000 are doing some social distancing to protect ourselves from the 1%. If we test and 20 people test positive (10 infectious and 10 not) then 980 people can return to their lives and only 20 need to intensify social distancing. The pandemic ends quickly.
We could cut the number quarantining even further by retesting using PCR and that’s good but not necessary. Also note that the 20 who test positive were already social distancing, albeit perhaps less carefully than ideal, so the additional cost is low and intensifying social distancing on the infectious reduces the transmission rate. I have ignored false negatives to focus on a key issue. False negatives will mean some transmission still occurs but that will be picked up by more frequent testing.
The takeaway is that when you are blind, you don’t need 20/20 vision to be much better off.
It’s taken far too long and it’s still not FDA approved for at-home use or for asymptomatic individuals but the new $5,15-minute, easy to use, Abbott test and the Trump administration’s promise to purchase 150 million of them is a big deal. Abbott has been building capacity for months according to their lead scientist interviewed in the Atlantic by Alex Madrigal and in a few weeks will be producing 50 million tests a month:
Madrigal: Fifty million tests a month is a huge number. That’s more than twice the number of tests the U.S. completes in a month. How did you ramp up production so massively?
Hackett: This was the challenge of this program. We needed some sort of reliable testing that could be affordable and that doesn’t require instrumentation. You need scale. The more frequently you could test people, frankly, even tests with lower sensitivity would be very effective at identifying people quickly and slowing the spread. As we were developing the test, there were people working in parallel looking at supply chain and logistics. Abbott took a lot of risk—hundreds of millions of dollars were spent building two new manufacturing facilities focused solely on those tests. We hoped we could come to a solution that would be where we needed it from an overall accuracy perspective, but if you weren’t building capability simultaneously, there was no way it could be the answer.
The US has performed about 80 million tests since the pandemic began, so an additional 50 million tests a month is a big increase in capacity. As noted, the test is not approved for at-home use but it’s a CLIA-waived test which means that a doctor’s office, a CVS or Walmart clinic, even a school nurse could qualify for a waiver and perform the tests. The test is not approved for asymptomatic individuals but I suspect that won’t mean much in practice, it can be prescribed off-label although the fact that a prescription is required is limiting. I hope the necessity for a prescription will be lifted as we get more experience with these tests. False positives (~1.5%) are low and by taking the strain off the PCR system we can improve triage and afford to do more double checks. False positives will be more of an issue as we wipe out the virus but that will take time.
I hope these tests will open up air travel within a month or two. I also hope to see more of these types of tests approved. Derek Lowe has more technical details.
It won’t be all smooth sailing, Abbott may not be able to produce as much or as quickly as they say they can and quality in the field may fall. The government may distribute the tests poorly. The virus could pickup in the fall, as in 1918. I expect more problems and challenges but we now have a chance to get ahead of the virus which is very welcome news.
Addendum: This type of public-private partnership with private firms building capacity in advance of approval for tests and vaccines on the foundation of government push and pull funding is exactly the structure that the Accelerating Health Technologies team has been recommending both to the US government and to governments around the world.
Jason Crawford at the Roots of Progress points out that we used to have a lot of ticker-tape parades. The most famous was perhaps the Victory Parade of World War II but we used to have many parades to celebrate technological and cultural milestones. There were huge celebrations, for example, when the final spike of the transcontinental railroad was nailed, when the Brooklyn Bridge was completed, and the Statute of Liberty dedicated. In the 1920s and 1930s there were big celebrations for aviation pioneers including for Charles Lindburgh, Amelia Earhart and Howard Hughes and that tradition continued in the 1960s and 1970s with multiple parades for the astronauts:
During the early space program, there were also several NYC ticker-tape parades for astronauts—not just the Apollo 11 heroes, who went on a world tour after the Moon landing, but missions before and after as well:
- 1962, March 1 – John Glenn, following the Mercury-Atlas 6 mission.
- 1962, June 5 – Scott Carpenter, following the Mercury 7 mission.
- 1963, May 22 – Gordon Cooper, following the Mercury 9 mission.
- 1965, March 29 – Virgil “Gus” Grissom and John Young, following the Gemini 3 mission.
- 1969, January 10 – Frank Borman, James A. Lovell, and William A. Anders, following the Apollo 8 mission to the Moon.
- 1969, August 13 – Neil Armstrong, Buzz Aldrin, and Michael Collins, following Apollo 11 mission to the Moon.
- 1971, March 8 – Alan Shepard, Edgar Mitchell, and Stuart Roosa, following Apollo 14 mission to the Moon.
- 1971, August 24 – David Scott, James Irwin, and Alfred Worden, following Apollo 15 mission to the Moon.
One of the last big ticker tape parades was in November of 1998 for John Glenn and the astronauts of Space Shuttle Discovery but since then the number of such parades has declined.Why? Has the number of accomplishments worthy of a parade declined? Or have we become complacent or even cynical about progress?
Jonas Salk famously turned down a ticker tape parade for the creation of the Polio vaccine but there was excitement and celebration around the world. When the time comes, I hope that we will enthusiastically celebrate science and the success of a COVID vaccine.
If you think the FDA has been slow at approving new coronavirus tests just look at their process for approving sunscreen products.
According to EWG, the Environmental Working Group, the FDA has been too slow to test old ingredients for safety and too slow to allow new ingredients on the market thus leaving us with sunscreen products which are neither as safe nor as effective as they should be. In particular, Europe has better sunscreen protection than the United States. Here’s EWG:
Americans have fewer choices and notably poorer protection than Europeans do from ultraviolet A rays in their sunscreen options. Although most U.S. sunscreens prevent sunburn effectively when used correctly, they aren’t as good as European sunscreens at preventing the more subtle skin damage produced by lower-energy UVA radiation. UVA rays have less energy and don’t burn the skin, but they can cause the skin to age, suppress the immune system and contribute to the development of melanoma.
…Between 2003 and 2010, sunscreen makers applied for FDA permission to use eight sun-filtering chemicals developed by European companies. Four of these – Tinosorb S, Tinosorb M, Mexoryl SX and Mexoryl XL – appear to be more effective than avobenzone, the most common UVA filter permitted by the FDA. The FDA’s failure to respond to these applications prompted Congress to pass the Sunscreen Innovation Act of 2014 (FDA 2014). This act requires the FDA to review new applications for sunscreen active ingredients within 300 days, but it doesn’t relax the standards companies must meet to prove new ingredients are both safe and effective.
In 2015, the FDA responded that the companies involved had not submitted enough information to prove their chemicals were, in fact, safe and effective for use (FDA 2015). The agency asked for more data, including complete study results, measurements of ingredient levels in people’s blood, and long-term studies on systemic toxicity and potential endocrine system disruption. The FDA has also proposed that all sunscreen ingredients, including those already in use, need to have adequate safety testing data.
Some information the FDA wants, such as complete copies of studies, might be easy for sunscreen makers to produce. But in other cases, the companies could take years to satisfy FDA requests. In the meantime, Americans are being shortchanged.
I first wrote about this issue in 2013 and seven years later, despite Congress passing a law in 2014, the FDA still has not acted.
My rule is very simple. I don’t think the FDA is better than the EMA so if any drug or device is approved in Europe it ought to be available for purchase in the United States with a label saying “Approved by the EMA. Not approved by the FDA.” (By the way, we do have reciprocity type agreements with Canada and New Zealand for food so this would not be unprecedented.)
Hat tip: John Thacker.
Addendum: You should actually get more sun to avoid vitamin D deficiency which is bad for a variety of reasons including, in my estimation, greater susceptibility to COVID.
Supply chains were hit hard early in the pandemic. Disinfectant couldn’t be produced because of a lack of bottles, tests couldn’t be processed because nasal swabs or PPE wasn’t available, the decline of passenger air traffic hit commercial delivery and so forth. I worry about forthcoming stresses on the vaccine supply chain. Billions of doses of vaccine will be demanded in the next year and a lot will depend on complicated supply lines including cold storage, air traffic, styrofoam, vials, bags, needles and many other inputs. Companies and the awesome team at CEPI (give them all a Nobel prize) are planning for vials and needles and other inputs but there are many non-obvious inputs higher up in the supply chain that also need shoring up.
Writing in Bloomberg, Scott Duke Kominers and I look at some of the odder inputs to vaccines like horseshoe crab blood, shark livers and the vaccinia capping enyzme, VCE. We are actually not too worried about horseshoe crab blood and shark livers as these are used in other industries. Shark livers, for example, are used to produce a lot of cosmetics so we should be able to divert supply as needed. VCE, however, is rarer.
DNA and mRNA vaccine technologies have shown promising results, and two of the leading vaccine contenders, from Pfizer Inc. and Moderna Inc., use mRNA technology. But mRNA has never been used to produce a commercial vaccine for humans, let alone at scale. And scaling these technologies may not be easy. In particular, mRNA degrades rapidly. To prevent this, it must be “capped” by a very rare substance called vaccinia capping enzyme.
Just over 10 pounds of this VCE is enough to produce a hundred million doses of an mRNA vaccine — but the current manufacturing processes for VCE require so much bioreactor capacity that making 10 pounds would cost about $1.4 billion. More important, global bioreactor capacity cannot support production at that level while also producing other vaccines and cancer-fighting drugs.
If we work hard now, we may be able to find more efficient means of producing VCE. Expanding bioreactor production and repurposing bioreactors from existing large-scale industrial applications will also help to lessen the pressure on the supply chains for multiple types of vaccines.
In addition to supply chains per se we also face the problem that companies are not raising prices enough. Ironically, this means that we need more public investment.
Of course, we might think that private companies would have incentives to coordinate supply chains themselves — and to some extent, they are doing so. But many have pledged to keep their vaccine prices close to costs, both out of altruism and because they may fear public backlash (or legal action) if they’re perceived as “price gouging” in the middle of a pandemic. And if companies don’t stand to profit much from Covid-19 vaccines, then they don’t have much incentive to invest in increasing capacity. In short: If prices can’t rise, then the only way to encourage companies to invest more in production is to reduce their costs — and that means we need public investment.
More generally, it’s not too late to be building more vaccine capacity and to repurpose bioreactor capacity from non-GMP sources, perhaps including veterinary and food sources. There are lots of vaccines in development. The science is promising. We need to take action now so that we can deliver on that promise.
Read the whole thing.
Social media are a coordination device and coordinated behavior has many advantages. Social media was used to motivate, organize and coordinate movements like the Arab Spring and Black Lives Matter. Of course, coordination can also lead to conspiracy theories like QAnon, twitter mobs that police political correctness and riots that lead to death and destruction. For better or worse, coordinated behavior is likely to increase, creating more and more quickly moving mobs. The use and abuse of such mobs is only just beginning. One insidiously clever prospect is the use of seemingly benign coordination to bring down a power grid. What if everyone turns on their air conditioner and lights at the same time? In How weaponizing disinformation can bring down a city’s power grid, Raman et al. discuss how such a scenario could be generate by something seemingly as simple as sending fake coupons!
Social media has made it possible to manipulate the masses via disinformation and fake news at an unprecedented scale. This is particularly alarming from a security perspective, as humans have proven to be one of the weakest links when protecting critical infrastructure in general, and the power grid in particular. Here, we consider an attack in which an adversary attempts to manipulate the behavior of energy consumers by sending fake discount notifications encouraging them to shift their consumption into the peak-demand period. Using Greater London as a case study, we show that such disinformation can indeed lead to unwitting consumers synchronizing their energy-usage patterns, and result in blackouts on a city-scale if the grid is heavily loaded. We then conduct surveys to assess the propensity of people to follow-through on such notifications and forward them to their friends. This allows us to model how the disinformation may propagate through social networks, potentially amplifying the attack impact. These findings demonstrate that in an era when disinformation can be weaponized, system vulnerabilities arise not only from the hardware and software of critical infrastructure, but also from the behavior of the consumers.
Hat tip: Kevin Lewis.
Addendum: California is once again issuing rolling blackouts. Welcome to the future.
It is sometimes said that economics does not predict. In fact, Lin et al. (2020) (SSRN) (including Colin Camerer) find that the classic supply and demand model predicts behavior and outcomes in the double oral auction experiment in thousands of different experiments across the world. The model predicts average prices, final prices, who buys, who sells, and the distribution of gains very well as Vernon Smith first showed in the 1960s.
Indeed, the results from simple competitive buyer-seller trading appear to be as close to a culturally universal, highly reproducible outcome as one is likely to get in social science about collective behavior. This bold claim is limited, of course, by the fact that all these data are high school and college students in classes in “WEIRD” societies (Henrich et al.,2010b). Given this apparent robustness, it would next be useful to establish if emergence of CE in small buyer-seller markets extends to small-scale societies, across the human life cycle, to adult psychopathology and cognitive deficit, and even to other species.
It is true that economic theory is less capable of explaining the process by which prices and quantities reach equilibrium levels. Adam Smith’s theory about how competitive equilibrium is reached (“as if by an invisible hand”) has been improved upon only modestly. The authors, however, are able to test several theories of market processes and find that zero-intelligence theories tend to do better, though not uniformly so, than theories requiring more strategic and forward thinking behavior by market participants. The double-oral auction is powerful because the market is intelligent even when the traders are not.
The authors also find that bargaining behavior in the ultimatum game is reproducible in thousands of experiments. Simple economic theory makes very poor predictions (offer and accept epsilon) in this model but the deviations are well known and reproducible around the world (participants, for example, are more likely to accept and to accept quickly a 50% split than a split at any other level).
The experiments were run using MobLab, the classroom app, and were run without monetary incentives.
Tyler and I use Vernon Smith’s experiments to explain the Supply-Demand model in our textbook, Modern Principles, and it’s always fun to run the same experiment in the classroom. I’ve done this many times and never failed to reach equilibrium!
Abstract: Can attitudes towards minorities, an important cultural trait, be changed? We show that the presence of African American soldiers in the U.K. during World War II reduced anti-minority prejudice, a result of the positive interactions which took place between soldiers and the local population. The change has been persistent: in locations in which more African American soldiers were posted there are fewer members of and voters for the U.K.’s leading far-right party, less implicit bias against blacks and fewer individuals professing racial prejudice, all measured around 2010. Our results point towards intergenerational transmission from parents to children as the most likely explanation.
from a new paper by David Schindler and Mark Westcott in ReStud. Black GI’s also experienced a society with much less segregation than in the United States.
Mixed race couples dancing in a London club, 1943. Original Publication: Picture Post – 1486 – Inside London’s Coloured Clubs – pub. 1943 (Photo by Felix Man/Picture Post/Getty Images)
From this review.
Here’s a good picture illustrating the difference between the PCR and Rapid Test. A PCR amplifies DNA and so if taken at the right time it will detect the virus before a rapid test will. But this happens when there isn’t much viral load and too little of the virus to be transmissible. Moreover, at these times, the virus is increasing rapidly so the rapid test will find the virus tomorrow. The PCR test will also pick up fragments after transmissiblity has passed which also isn’t very useful. A rapid test is very sensitive for doing what it is supposed to do, identifying periods of infectiousness.
Michael Mina has done a great job promoting rapid tests and I do think we are beginning to see some recognition of the difference between infected versus infectious and the importance of testing for the latter. What is frustrating is how long it has taken to get this point across. Paul Romer made all the key points in March! (Tyler and myself have also been pushing this view for a long time).
In particular, back in March, Paul showed that frequent was much more important than sensitive and he was calling for millions of tests a day. At the time, he was discounted for supposedly not focusing enough on false negatives, even though he showed that false negatives don’t matter very much for infection control. People also claimed that millions of tests a day was impossible (Reagents!, Swabs!, Bottlenecks!) and they weren’t impressed when Paul responded ‘throw some soft drink money at the problem and the market will solve it!’. Paul, however, has turned out be correct. We don’t have these tests yet but it is now clear that there is no technological or economic barrier to millions of tests a day.
Go yell at your member of Congress.
But people aren’t getting their tests back quickly enough.
Well, that’s just stupidity. The majority of all US tests are completely garbage, wasted. If you don’t care how late the date is and you reimburse at the same level, of course they’re going to take every customer. Because they are making ridiculous money, and it’s mostly rich people that are getting access to that. You have to have the reimbursement system pay a little bit extra for 24 hours, pay the normal fee for 48 hours, and pay nothing [if it isn’t done by then]. And they will fix it overnight.
Gates is correct. If companies were paid for speed they would increase capacity and move immediately to a stack processing (LIFO) model, as I described yesterday.
The whole interview is worth reading. Gates is restrained but you can tell he is angry. Bill has had it with the FDA, Trump, Mark Zukerberg, stupid anti-vaxxers like Robert Kennedy (who he was forced to listen to to get access to Trump), Congress and much more. I don’t blame him one bit. I am angry too.
Ideas Of India is Shruti Rajagopalan’s new podcast about India. This is going to be an excellent podcast, well worth subscribing to. Shruti’s first guest is Ajay Shah discussing his book with Vijay Kelkar, In Service of the Republic: The Art and Science of Economic Policy. As you may recall, I called In Service, the new Arthashastra, the book every policy maker and future policy maker should be given while being told, “before you do anything, read this!”
Here’s one bit from Ajay in the podcast:
[S]tate capacity is very hard to change. It evolved very slowly, but it is something you learn. There’s a learning by doing for a republic to learn to achieve state capacity. So we would tell a more constructive story of saying,” Pick a few battles, do a few things, learn how to do them well.” Then maybe in the future you might like to creep out, while understanding that these are 20-year, 40-year, 80-year, hundred-year journeys. Don’t think that these things can be solved in two years.
…There’s a quote in the book from Kaushik Basu where he said that we have libertarianism of necessity, and we have libertarianism of choice. In India, we have to do libertarianism of necessity because we every day confront the malfunctioning state institutions. We’ve always got to think, can this work? Would it go wrong? We’re surrounded by unchecked coercive power in the hands of very frail state institutions, and that creates limits on state capacity. So I think that’s the way our lived experience in India has brought us.
Exactly right and very consistent with the argument that Rajagopalan and I make in Premature Imitation and India’s Flailing State:
In the alternative view put forward here…presumptive laissez-faire is the optimal form of government for states with limited capacity and also the optimal learning environment for states to grow capacity.
A COVID test that doesn’t come back in a few days is close to useless and PCR tests are taking a long time to process:
NYTimes: Most people who are tested for the virus do not receive results within the 24 to 48 hours recommended by public health experts to effectively stall the virus’s spread and quickly conduct contact tracing, according to a new national survey by researchers from Harvard University, Northeastern University, Northwestern University and Rutgers University….People who had been tested for the virus in July reported an average wait time of about four days. That is about the same wait time for those who reported taking a test in April. Over all, about 10 percent of people reported waiting 10 days or more.
…“A test result that comes back in seven or eight days is worthless for everybody — it shouldn’t even be counted,” said Dr. Amesh Adalja, a senior scholar at the Johns Hopkins University Center for Health Security and a physician in Pittsburgh. “It’s not a test in any kind of effective manner because it’s not actionable.”
One seemingly severe but potential solution is to change how tests are processed. Right now it’s mostly first come, first-served but this means we can easily have a situation where everyone eventually gets a test result but all the results are useless because they take a week or more to process. I propose instead that any test that can’t be reported back in 3-4 days be thrown out immediately. Labs should focus only on processing tests that can be reported back quickly.
One way of thinking about this is to use a stack or last-in first-out (LIFO) model for testing. In a stack model the newest test request is pushed onto the top of the stack and the next test to be processed is popped off the top of the stack. One disadvantage of this model is that some test requests will never be processed (they should be removed from the bottom of the stack and returned as null results). Some people will be angry.
But the stack model of testing has a huge advantage over first-come, first-served. Namely, just as many tests will be completed as under the current model but the tests results will all come back faster and be much more useful. What would you rather have, guaranteed stale test results or fresh results with some possibility of a null return? Since a stale result is not much better than a null it seems obvious that the stack system is superior. Most importantly, faster, more useful tests will help to end the crisis by reducing the number of infections.
Addendum: See also my posts Pooled Testing is Super-Beneficial and Frequent, Fast, and Cheap is Better than Sensitive on other methods to improve testing.
On August 2, bio-statistician Steven Salzberg argued that We Should Consider Starting Covid-19 Vaccinations Now. But, under immense pushback, including an article by another bio-statistician Natalie Dean writing in the NYTimes, he changed his mind and reversed course. I was frustrated by both sides of the debate since neither “biostatistician” presented any numbers to justify their arguments! So let’s do this better.
Suppose you take a vaccine now as opposed to (optimistically) on Dec. 1, 2020. From May 1 to August 5 we averaged 1001 deaths a day. There are 117 days between now and Dec 1 so at that rate there will be ~117,000 additional deaths by Dec. 1. Let’s call it 100,000. There are 324 million people living in the United States so the probability of dying from COVID in the next 117 days is 1/3240 or .03%.
Now what are the risks of dying from a vaccine? We don’t know these risks but suppose the vaccine is given to 100 million people in the United States then in order for there to be an equal number of deaths the probability of death from the vaccine would have to be 1/1000. That’s unlikely but not impossible!
Furthermore, phase three trials are the acid test for efficacy. Results from many phase II trials look good but we will learn more in a larger, more varied population actually at risk for the disease. We will also will learn which vaccines are better, e.g. Novavax’s protein based vaccine looks much better than others in early trials and that will become clearer with larger trials.
Overall, the numbers here do not make a strong case for vaccinating early. I’ve long argued that the FDA is much too risk averse in approving new drugs but vaccines are meant to be given to large numbers of healthy people which makes risk aversion more reasonable.
Note, however, that these numbers are for a randomly chosen member of the population but the people choosing to vaccinate early will not be randomly chosen. If you are an African-American or Latino, for example, your risks are higher. Your risks are higher still if you are an older, male, African-American or Latino physician, nurse, taxi driver or nursing home resident. In these cases, my judgment is that the benefits swing towards early vaccination. The benefits would be larger still if we assume that a vaccine won’t be available until 2021.
I’ve focused on deaths. Clearly, there are also other health risks but they fall on both sides of the equation.
A mass vaccination campaign in advance of phase three clinical trials would be unwarranted. Vaccinating large numbers of healthy people has real risks. Nevertheless, in my view it would not be unreasonable for someone at high-risk of COVID to choose to be vaccinated before waiting for longer clinical trials and such early vaccination, as Tyler noted, would also provide valuable information for everyone else.
Addendum: The Open Source RADVAC vaccine is one option for those with the requisite medical expertise.