Among the many consequences of the Covid-19 pandemic, and the measures taken to control it, there has been an epidemic of whataboutery. The starting point is the claim “we have locked down the entire economy to reduce the number of deaths from Covid-19, but we tolerate comparably large numbers of deaths from X”. Popular candidates for X include smoking, road crashes and influenza. In most, though not all, cases, the inference is that we should accept more deaths from the pandemic. Indeed, the majority of those using this argument are also opposed to any proposal to do more about the various examples of X they cite
I’m going to take the contrapositive, and argue that the inconsistency pointed out here should be resolved by taking stronger action to reduce avoidable deaths from a wide range of causes, with the primary examples being road deaths and smoking.
While whataboutery on these topics typically suggests that society has made a decision to tolerate deaths from these causes, the reality is that there have been increasingly stringent measures to reduce them, adopted over many years, and that in both cases, the ultimate objective (explicit in some jurisdictions, implicit in others) is to reduce deaths to zero. In the case of roads, this aim is expressed in Vision Zero, adopted initially in Sweden and subsequently in a variety of other places. The UK government aims to end smoking by 2030, and most governments have interim targets which imply ultimate elimination of smoking.
With or without explicit targets, the policy approach everywhere has been much the same. Restrictions aimed at reducing the risk in question have been introduced gradually over many years, with each new restriction providing a starting point for the next. In Australia’, for example, partial bans on tobacco advertising were introduced in the late 1980s. These were followed by complete ad bans, then by compulsory health warnings in small print, and finally by a requirement that cigarette packets should display gruesome photos of the consequences of smoking. At the same time, from an initial situation where smoking was universal, it has been progressively restricted in all public spaces, and where children may be exposed (as in private cars).
There is indeed an inconsistency here. If the restrictions in place now are justified in terms of a balance between health costs, damage to non-smokers and the restrictions on the rights of smokers, they would have been even more justifed 30 or 50 years ago, when the damage done by smoking was much greater. Coming back to Covid whataboutery, the inconsistency is not between accepting deaths from one source and not another, it’s between the urgent action necessitated by the pandemic and the slow pace adopted in other cases.
The slowness with which policies aimed at ending smoking, or road deaths, is easily explained. Governments have introduced them at a pace that avoids substantial political costs, and the risk of sustained non-compliance. In the case of smoking, for example, it is necessary to deal both with powerful and unscrupulous tobacco companies, using every available tool[1] to resist controls, and with a large addicted population, some (though not all) of whom have no desire to quit.
The success (so far) of lockdowns in controlling Covid, and their general acceptance outside the US, suggests that we should move more rapidly to eliminate public health risks, even where this involves coercive measures to stop people endangering others, and to prevent young people from endangering themselves. For example, partial bans on smoking in public places, or in the presence of children, should be made total. A more ambitious proposal of this kind would be to raise the smoking age, one year at a time, so that young people currently under the legal age would not be allowed to smoke until they were, say, 25 (hardly anyone begins smoking as a mature age adult, which is in itself an indication that it is not a choice open to a rational defence).
In the case of road deaths, the most obvious measures are lower speed limits in urban ares, and a greater willingness to take dangerous drivers off the road permanently. These measures will be adopted eventually – the only question is how many innocent lives will be lost before they are.
fn1. The tobacco companies not only lobbied directly, and funded a variety of front groups (astroturf smokers rights groups and free-market think tanks), but fought Australia’s packaging laws through international trade actions, ginned up by bribing governments or exploiting the Investor-State Dispute Settlement clauses of trade agreements. They were defeated, but almost certainly succeeded in deterring poorer countries, which could not afford such fights, from following Australia’s lead.
John Quiggin 
Me thinks a more approriate title would be “the case for more nanny-stateness” but I know I’m in the Robinson Crusoe minority here 🙂
One person’s modus ponens is another person’s modus tollens
With cars the solution might be to require anti-collision and anti-pedestrian splat technology . When every car has to have sensors and the ability to respond appropriately if the driver doesn’t then it’s not much of a step to full self driving, which — when done by Waymo — currently looks much safer than letting humans drive and will presumably will improve in safety each year. I expect the number of humans driving will rapidly fall, but the holdouts can eventually be required to retrofit safety equipment if they insist on driving antique idiot cars around. (I have no problem offering people free robo taxi service in return for giving up a dangerous vehicle.)
I think a strong component of this is status quo bias followed up with cherry picking.
Cars will always need a driver:
1] to take responsibility,
2] automatic systems are re-active, but can not be predictive.
As I mentioned on CT, these whataboutery-ers simply don’t know anything about the numbers of deaths involved. They want us to listen to their risk assessment but they know nothing about the risks involved. It’s not some kind of logical gotcha, it’s just ignorance.
@1 If your best argument is a silly pejorative, you ought to think about why you are in the minority.
“the most obvious measures are lower speed limits in urban ares”
What does the research say – if you lower the speed limit by x kph you reduce deaths by how many?
I dunno John, he has a point that we do need more nanny-stateness. For starters a rolling age limit on smoking, so for anyone who is not already over the legal smoking age, smoking is always illegal. Secondly, a reduction in the speed limit so that more fatalities can be prevented. Thirdly, ever-improving clean air standards on cars with the ultimate goal of eliminating all non-electric cars as quickly as possible, followed by a ban on the use of the internal combustion engine and all forms of carbon-emitting fuel.
A little more nanny-stating would do the world a lot of good!
John Homan: have you studied the AI algorithms making up say Tesla’s autopilot?
Ronald: the reason we can be quite sure self-driving cars will be safer is the strong human bias about agency in assessing risk. One plane or train crash a year? (Typically zero.) Eek! 300 or 3000 (USA) killed in car crashes? Not news. Full self- driving will have to be 10 or 100 times safer than Joe Driver before it’s approved. They’ll get there.
With cars the solution might be to require anti-collision and anti-pedestrian splat technology .
Why not go with the solutions we already have that are known to work? Hoping that in the future technology might save us is very much accepting that we’ll keep on killing people unnecessarily in the meantime. Much like leaving the pandemic problem until someone else makes a vaccine available… even the UK bottled that idea.
The known solutions are Copenhagen as the easy first step followed by The Netherlands once you have the basic ideas accepted. In short the solution is to bicylate the parts are are too spread out to pedestrianise (youtube link, channel worth looking at) and make public transport the dominant motor vehicles. Yes, they’re both small countries with high population density… much like the inhabited parts of Australia but if you want distance there’s the option of a quick train trip to Paris (3 hrs) or even Barcelona (9 hrs).
Sorry, I stuffed up the links there. Can I have another go? (JQ, please feel free to delete the broken dupe)
With cars the solution might be to require anti-collision and anti-pedestrian splat technology .
Why not go with the solutions we already have that are known to work? Hoping that in the future technology might save us is very much accepting that we’ll keep on killing people unnecessarily in the meantime. Much like leaving the pandemic problem until someone else makes a vaccine available… even the UK bottled that idea.
The known solutions are Copenhagen as the easy first step followed by The Netherlands once you have the basic ideas accepted. In short the solution is to bicylate the parts are are too spread out to pedestrianise (youtube link, channel worth looking at) and make public transport the dominant motor vehicles. Yes, they’re both small countries with high population density… much like the inhabited parts of Australia but if you want distance there’s the option of a quick train trip to Paris (3 hrs) or even Barcelona (9 hrs).
There was a lockdown in Wuhan and possibly elsewhere in China but there was no comparable lockdown in Australia. There were and still are merely restrictions on people’s behaviour in public, setting aside mandatory quarantine in hotels at public expense. There are and have been restrictions on people’s behaviour on public roads via a mechanism called traffic rules, including traffic lights. There have been restrictions on smoking in public places for a long time. In some municipalities the only area where smoking is not prohibited is on the edge of a major road where exhaust fumes from heavy traffic dilute the cigarette smoke (this seems to be the underlying reasoning of those who make the rules), and not in front of public buildings such as schools and hospitals (where exhaust fumes apparently are assumed not to affect the protected people).
As is the case with the corona virus restrictions in place at present and expected to persist in one way or another until and if a vaccine becomes available, the mechanisms regarding road usage and smoking to protect the public health is not expected to work without errors. That is, further corona virus infections are expected, some involving serious illness and even death. Similarly, road accidents are expected to happen, some involving serious injury and even death. Similarly, smoking is expected to involve some serious illnesses and even death.
There are however differences between the treatment of smoking and the other two examples of governmental measures taken to protect public health. Corona virus measures and road traffic rules involve public places, not individuals’ private homes. By all means promote total smoking ban in public places.
This leaves the taxation question.
It is easy to find data on how excise on tobacco products enter the government’s revenue estimation:
https://treasury.gov.au/sites/default/files/2019-07/t392185_tp_tobacco_forecasting_working_paper.pdf (“The paper”)
“Tobacco excise as part of the Commonwealth tax mix
The significance of tobacco excise to the Commonwealth tax system is determined by the balance
between declining clearances and increasing excise rates. Given there has been a strong trend decline
in tobacco consumption per capita over the last four decades (see Chart 3), tobacco excise increasing as a percentage of GDP is a result of the excise rate increasing enough to offset the lower consumption.
“The period from 2000-01 to 2008-09 contained no excise rate increases beyond indexation, and with
the exception of 2002-03, tobacco excise declined as a percentage of GDP in every year over that time.
2010-11 was the first full year after the 25 per cent excise rate increase in May 2010, and saw a
significant increase in excise collections. The next excise rate increase above the indexation rate was in
2013-14. Tobacco excise has increased as a percentage of GDP in every year since then, and is forecast
to continue to do so until the last scheduled non-indexation increase in 2020-21. Absent further large
excise rate increases, it would be expected that tobacco excise would decline as a percentage of GDP
beyond 2020-21, albeit at a slower rate than from 2002-03 to 2009-10 because of indexation to wages
rather than prices.” [Source: The paper, pp 10-11]
It is much more difficult to find sensible numbers for the health costs of smoking. A paper by Whetton et. al., titled “Identifying the Social Cost of Tabacco Use in Australia in 2015/16”, National Drug Research Institute (NDRI) Curtin University, May 2019, and funded by a government research grant, arrived at a headline number of $136.9 billion. http://ndri.curtin.edu.au/NDRI/media/documents/publications/T273.pdf
I say headline number because this is the number that appears when using a google search and it is repeated in various places.
Lets reflect on Whetton et. al. social cost number of $136.9 billion in 2015/16.
The estimated cost of the corona virus pandemic to the Australian health system was estimated to reach $1 billion (SBS news, March 2020).
The Government has allocated more than $230 billion in stimulus measures, including the $130 billion JobKeeper wage subsidy, in response to the coronavirus.[Source: ABC news, 3 days ago).
Whetton et. al. ask the reader to accept that the smokers in Australia generate at least as much costs to the Australian public as the entire JobKeeper wage subsidy and the cost to the health system thrown in.
How did Whetton et al. arrive at this astronomical number of $136.9 billion?
As can be checked on the above link, out of these $136.9 billion only $1.7 billion are health costs that are somewhat comparable to the health costs estimates regarding the corona virus. That is, hospital admissions.
To make the numbers at least a little bit comparable, the cost of cruise ship tickets should be added to the social cost of corona virus because Whetton et al add the cost of cigarettes.
Although the practitioners of high risk sport, which require rescue operations and hospitalisations as well as all types of air polluters, do not pay a special tax, I am confident in saying (as a smoker) smokers wouldn’t mind paying say $2 billion excise in addition to GST p.a. for hospitalisation costs. But this offer would not square with the Treasurer’s revenue plans, would it?
The entire argument about smoking and health ignores the behavioural response of smokers to the information about the negative health effects of smoking. For example, there has been little if any objection by smokers to refrain from polluting other people’s atmosphere in public places when the danger of second hand smoke became known (just like people having in the main followed the advice on social distancing in the current pandemic). Furthermore, even in their private homes, many smokers do not smoke indoors any more – very difficult for people who live in apartment towers.
Social cost estimates as contained in the Whitton paper remind me of the travel time saving benefit estimates in cost benefit studies used in the decision making process for roads. The former leads to an excuse for raising excise taxes while the latter leads to excessive road construction and air pollution that is harmful to health.
PS John Homan: this arXiv pdf on self-driving AI is pretty incomprehensible, but it has a lot of references to prediction, none of them suggesting it can’t be done. https://arxiv.org/pdf/1910.07738.pdf
Example: “Provided by Daimler AG R&D and University of Amsterdam, this dataset fits the topics of pedestrian detection, classification, segmentation and path prediction.” My italics. I imagine the software assesses which way pedestrians are facing and predicts they will either move forward in the same direction at normal walking speed, or not move at all. It’s technically possible to walk backwards into oncoming traffic, but so improbable that it can probably be ignored. The same goes for accelerating like Usain Bolt off the blocks. Perhaps the AI just makes these into rebuttable Bayesian presumptions, ask the authors not me.
John Homan,
The device you used to leave your comment has predictive text.
Moz, I am of the opinion that people can have cars if they like as long as they don’t kill people either directly or indirectly through pollution and environmental damage and pay for roads. There’s no reason why cycling and public transport can increase at the same time this is worked towards. Although it may not happen if people are lazy/despise the presence of their fellow human beings. Once the majority of vehicles are self driving the distinction between private transport and public transport is likely to blur in any sensible system. Note that if self driving cars are allowed to circle the block while their owner gets a haircut that’s not a sensible system.
How about us throwing away our freedoms and privacy ,living in fear, trashing our international goodwill ,helping destroy the Middle East ,creating a refugee crisis ,wasting billions of dollars and starting a war on Islam because a few stupid and/or mentally ill kids killed or injured a couple of Aussies ? Every time a terrorist tries to hurt one person, or even allegedly thinks about it ,we go into a de facto state of national emergency and let the government do whatever they like.
One idea for reducing deaths and injuries from road accidents would be to investigate them properly. By this I mean replacing the traditional model for road, rail and marine accidents, designed to allocate moral blame and legal liability (civil and criminal). Instead emulate the model for aviation, which seeks instead to find out how the accident could have been prevented, taking a systems approach to both mechanical and human aspects. Air crash investigations are lengthy, thorough, independent and expensive. (The pilot was drunk? Sure, that was the proximate cause. This is where current road and marine investigations stop, and aviation ones continue. Why did the pilot drink? Work or domestic stress? Did HR not spot the problem? Why did other crew members not intervene? Etc etc.) But they have made deaths in civilian commercial flying a great rarity. The number of road accidents is so large that initially only a sample could be treated this way.
I doubt it’s a good idea to impose criminal penalties on the people who are doing the smoking. It would be a better idea to impose criminal penalties for selling tobacco products to people below a stipulated age and having that age increase by one each year.
James Wimberley @ 2:28 am
“One idea for reducing deaths and injuries from road accidents would be to investigate them properly… “taking a systems approach to both mechanical and human aspects”…
James, thanks for posting this, it is such a good idea. I would vote for such in policy. I knew and supported a gifted system dynamics modeller in the US, who’s son took his life, and tried in vien to get this idea recognised. And failed.
And further extend this to suicide, ptsd, and many other stigmatised and ‘against the zeitgeist” / “costs” socially / politically / religiosly unpalatable human conditions.
And organisational culture problems as “”What was it about the culture of this particular place that that failed?” Professor Duckett said.”
https://abc.net.au/news/2020-05-17/maryborough-hospital-stillbirth-prompts-review-maternity-service/12239458
Any ideas as to how to embed this idea into the polity? Via a Cost benefit? If done by JQ and- Richard Holden & Bruce Preston: “The costs of the shutdown are overestimated – they’re outweighed by its $1 trillion benefit”.
https://theconversation.com/the-costs-of-the-shutdown-are-overestimated-theyre-outweighed-by-its-1-trillion-benefit-138303
We sorely need such an analysis “taking a systems approach to both mechanical and human aspects” for the post covid world.
Not a bad idea, James, but obviously a lot easier to implement if humans aren’t permitted to drive. (Of course, they can pretend drive if there is a suitably safe computer and sensor system ready to take over then they do something dangerous. A self driving car would be a great place for people who want to improve their driving skills to do so.)
Perhaps we should make it a law that if you purchase tobacco you’re not allowed to buy other drugs that are more fun?
However, we may need to develop these drugs, as there are generally drawbacks to pretty much all “fun” drugs at the moment. But rather than improving chemistry the best best for improving the utility of recreational drugs may be to improve monitoring and dosage. Many people report that a certain level of alcohol intoxication can be enjoyable but many appear to lack the ability to reach that level and stay there. If a person was going to exceed the level that results in the most enjoyment they could be charged more for a drink that would take them over that level. Since they would not be too far gone at this point they should still be able to respond to a price signal.
J-D it is illegal already to sell cigarettes to people under the age limit, and it’s illegal to smoke under the age limit. It works fine. The main effect is that people start smoking later. So increasing the age limit would simply change the number of people to whom the law applies. And making it a rolling increase means that no one yet to come of age will be suddenly declared to be breaking the law. It’s a smooth, easily enforcable way of criminalizing smoking without actually criminalizing anyone.
Since the sandpit has been taken over by conspiracy theories and has become Graeme Bird’s squawkbox, I’ll dump this information here – I saw on Twitter a new study released in the journal Cell that shows a) immunity developing in people who have recovered from the virus and b) the possibility of cross-immunity from other coronaviruses, which raises the possibility that 40-60% of the population are already immune or partially immune due to having recent colds. If so that raises some very scary questions about this virus’s infectiousness in a naive population (half hte population not susceptible suggests its R0 in a naive population would be close to 10). It also suggests that new infections could surge in summer, when the effect of past colds begins to wear off.
I haven’t assessed the study itself because I know nothing about immunology, but it’s interesting. Also the idea that infection with a common cold coronavirus could confer immunity raises the possibility of a simple immunization program, though it wouldn’t be without its side effects …
Faust, a less dangerous coronavirus that acts in a way similar to how cowpox protects against smallpox could be useful to have, but is problematic. While I might be willing to infect myself with it to protect others from COVID-19, if we expose elderly people to it it might kill 1% or more of them, in much the same way colds kill a significant number of elderly people every year. So in the right circumstances it might be great, maybe the US and UK, but probably of no use in Australia at this point.
Ronald: Once the majority of vehicles are self driving
I kind of like that idea, mostly because so few people can afford to spend $AU100,000 or more on a new car. If we made those mandatory we’d have much safer roads. But, like the idea of making motorists pay the actual cost they inflict on society, that is a dream.
There are arguments that current self driving technology would be safer than current people drivers, but that rests entirely on the experiment not having been done. Self driving is still at the “a phenomenally expensive collection of sensors and computers can sort of drive in many situations if properly maintained and constantly supervised”. I expect that on motorways, and at low speeds in motorway-type urban roads automated trucks will be very safe, much safer than existing trucks. But, and this is key, they will be so because they will be expensive new trucks only used by fleet operators with deep pockets.
IMO the question is not “is the new fighter jet equivalent car” safe, but “is a 10 year old self-driving car with the more expensive/fragile sensors bypassed”, still safer? Because that’s a much more accurate reflection of actual road conditions than anything we’ve seen from the current generation. They *wash* those cars every day for crying out loud. They don’t drive for a month with the “clean LIDAR 4 dome” siren blaring helplessly into a blob of silicone sealant. But even so, Telsa are doing a wonderful job of showing us just how far self-driving technology has to go before it’s even vaguely usable by the average motorist. People buy the “autopilot” thing, ignore the limitations, and every now and then one of them dies in a remarkably stupid way.
ronald indeed. Although if it has already infected 40% of the population and the required threshold for herd immunity is 50%, we could get there by infecting a few more healthy adults. (The required threshold is much higher than that, so it’s irrelevant). It’s all speculation anyway at this stage.
I guess you’re right, Moz. I’ll go tell the boys down at Bell Labs to forget about this transistor thing they’re working on. I have to admit it was a pipe dream to think we could do better than vacuum tubes. And now I’ve got to sweep up the chads from all the punch cards this email required.
I’m guesstimating America will have circa 300,000 Covid-19 deaths by November’s presidential election. I can see the Republicans engaging in quite a bit of otherworldly whataboutery to make this particular warthog look acceptable.
However it seems that 40% of US voters (those who turn on the day rather than potential voters) are rusted on Republicans for whom even 500,000 Covid-19 deaths would make no difference to their vote. All Trump and co. have to do to win their vote is keep pumping out #Obamagate, #Deepstate #QAnon etc nonsense and the faithful will lap it up. Plus you have thousands of hugely influential white evangelical preachers telling the faithful to vote Trump as he is the anointed one #PraiseJesus. It’s amazing that democracy even works given the craziness of so many voters.
Faustusnotes: preliminary results from a very large antibody study in Span (n = 61,000), which has had 27,000 deaths, show a 5% population infection rate. This translates to a 1.15% IFR, in the consensus range. (https://www.eurasiareview.com/15052020-spain-5-of-population-developed-antibodies-against-covid-19/ ; press release in Spanish https://www.newtral.es/wp-content/uploads/2020/05/ENE-COVID_Informe-preliminar-cierre-de-la-primera-ronda_13Mayo2020-1.pdf)
This is much more reliable than the smaller studies showing the very high immunity rates you are relying on. For Spain, a back-of-an-envelope calculation suggests that herd immunity of 60% would require (43m. x 60% x 1.15%) = 297,000 deaths, over ten times the current toll. Herd immunity by infection not vaccination is morally odious (since there is a dynamic TTI alternative) and politically impossible. If Trump gets his way about reopening the US economy before the pandemic is contained, regardless, he will be forced to flee into exile like Idi Amin.
I think you’re misunderstanding me James. I am referring to a study in Cell which suggests people who have previously had a different coronavirus (i.e. the common cold) could have immunity to SARS-2. This leads to an estimate of 40-60% immune due to prevalence of common cold, not that 40-60% of the population have already had SARS2.
i saw reports on the Spanish study and it sounds right to me – 10x as many people infected as reported, which I’ve been assuming from the beginning. It also indicates a very high IFR, as you say, and shows just how insane the UK’s herd immunity strategy would have been had it been allowed to continue.
(I’ve been saying all along that herd immunity by infection is a terrible idea but I don’t have your faith in the voting public’s response to its operation. The UK has the second worst infection in the world and still 39% of people polled approve of Johnson’s handling of it. There should be no one in the UK who thinks what has happened there is acceptable).
Wikipedia clams that only 15% of common colds are produced by coronaviruses rather than the more common rhinoviruses. I don’t see how 60% herd immunity from that pathway is possible. I’d need a lot more confirmation of the Cell article before taking this seriously.
One nugget in the Spanish study may be relevant here. 13% of those who had PCR-tested positive for the covid virus had no detectable antibodies to it. They must have recovered some other way: possibly existing antibodies to a closely related virus. Possibly also something completely different. Yeast defeats its own viruses without an immune memory at all. https://phys.org/news/2016-10-yeast-gene-rapidly-evolves-viruses.html
PS: Johnson is badly under water in the polls now, with a death toll of 34,500. The same envelope calculation as I did for Spain would give a herd immunity death toll of (66m x 60% x 1.15%) = 460,000. I maintain that such totals are qualitatively different to what we have seen so far. They would not only destroy the governments responsible but threaten constitutions, with Scottish and Catalan secessions a near-certainty.
Hugo: I agree with your ballpark of 300,000 US deaths. Simply scaling up current reported Spanish deaths by population gives 201,000 US deaths. Spain has universal health care, a better safety net, negligible libertarian cranks, and a rational central government, so a 50% Trump/Santis premium looks about right.
James I’m not sure where I got the 40-60% from. The wikipedia number and the Spanish results seem consistent though.
In addition to your reasons for ballpark 300k US deaths, let’s note that death counting isn’t complete yet and every other country has found its confirmed covid death counts to be a large underestimate.
However, I think your ballpark is going to be very wrong, because cases are about to begin booming again in the US and it’s going to continue to see 25,000 or more cases a day for the foreseeable future. Right now Texas, Wisconsin, Alabama, Florida, Minnesota, North Carolina, North Dakota, Oklahoma, South Carolina, and Virginia are seeing sudden large increases in the last few days, and their reopening has just begun. Meanwhile cases are not going down in a bunch of high-burden states like California, Pennsylvania, Washington and New Jersey, and are stable in most other states. Ohio for example is still seeing more cases a day than Sweden with no decline in numbers. With the reopening we can only see these increase further.
It’s madness to reopen with these kinds of numbers. The UK has 3500 cases a day and is beginning to reopen, if this leads to an R of just 2 – half this virus’s potential – they’ll be back to the growth rate from the peak of their epidemic in a week. Some of these US states (like Wisconsin) are doing a completely uncontrolled reopening when they have daily case numbers in the hundreds. The results of that are going to be very bad, and your death estimate is going to be wrong by a factor of 10.
Ronald, there is an enormous gulf between “all progress is stupid” and “completely automating the operation of motor vehicles is an idea without flaws”. I just can’t see this as “so obvious no debate is necessary”.
I have written software that could kill people and it was not easy, even in a very constrained environment where pressing stop was always safe. Doing so for a large set of different environments, most of which contain unpredictable objects and where pressing stop can kill… that is something I think we need to be very careful of. Doing it where one obvious consequence of doing it perfectly is that people die… that’s even more difficult.
There’s a whole lot of trade-offs to be made that even people can’t agree on, so writing software to make those decisions is pretty hard.
I *can* see that glibertarian tech bros don’t want to have those debates, and would rather impose their philosophy on everyone else. They’re doing that now and it’s working for them. But frankly, Elon Musk saying “let the peasants die, I need more money” makes me think he shouldn’t be allowed anything to do with self-driving cars.
Faustus,
James is simply agreeing with my ballpark estimate of 300K USA Covid-19 deaths by the USA presidential election, on 3 November 2020. I think 3 million deaths by that date is a tad far fetched.
Anyway, time will tell.
Meanwhile, Trump’s obsessive whatabout Obama tweets are ramping up again. Many of Trump’s supporters literally think Obama is some type of anti-Christ who is the puppeteer for the #deepstate and everything is his fault. America seems too weird to be real at times.
Hugo, if the epidemic there were to fade away by November at the current rate I would agree with you, but that’s not going to happen, so I think you’re underestimating. But yes upon reflection you’re right, 3 million deaths will only happen once the whole country is infected, which probably won’t happen. So maybe you’ll only be wrong by a factor of 2-5!
Which law are you looking at? The Public Health (Tobacco) Act 2008 (NSW) prohibits the first of those (and imposes criminal penalties for breach of the prohibition) but not the second.
Well there you go then J-D, there’s nothing to stop us increasing the age limit and criminalizing the sale but not the purchase, as you suggest. (Sorry I was wrong about the second)
JQ said “The starting point is the claim “we have locked down the entire economy to reduce the number of deaths from Covid-19, but we tolerate comparably large numbers of deaths from X”.
Which is greater – fear, terror, death awareness or freedom?
“Coping with ‘Death Awareness’ in the COVID-19 Era
According to terror management theory, people can have surprising reactions
…
“Findings from psychology help to explain these polar-opposite reactions—and how we can follow our best instincts rather than our worst ones. It all seems to come down to our terror about death. As cultural anthropologist Ernest Becker suggested in 1973, our capacity to reflect on our self poses a problem for human beings: the awareness of the existence of the self signifies that it will one day cease to exist. Within psychology, terror management theory studies how we react when death is made salient to us. In their book The Worm at the Core, Sheldon Solomon and his colleagues describe how terror management theory starts with the assumption that, like other living organisms, human beings have an instinct for self-preservation and survival. But unlike other organisms, our intellectual capacities make us painfully aware that one day we will die.”
https://www.scientificamerican.com/article/coping-with-death-awareness-in-the-covid-19-era/
In Australia if you buy a new car and drive it until the end of its life, accidents will cost you an average of around $33,000. This includes damage to the vehicle and not just to people, sign posts, dogs, etc. New cars, electric ones, are or will be drive by wire. The additional equipment cost of making them self driving is a couple thousand dollars, although it will depend on what’s finally required. I expect the cost to rise as Mister Magoo sensors are given up on and money thrown at improved hardware and computing power. But then the cost will fall a long way. If self driving cars have half the accident rate then that will be worth about $16,500 over its lifespan. (More actually, since it would, on average, have a longer lifespan.) So a fairly small equipment cost can easily pay for itself and lower the cost of driving. It also saves of human beings, if that’s something you care about.
KT2 says: “We sorely need such an analysis “taking a systems approach to both mechanical and human aspects” for the post covid world.”
A systems approach to analysing the relationship between behavioural restriction measures and the economy, as represented by GDP, has been carried out by Michael Meyer-Hermann,[1] Head of the Department Systems Immunology at the Helmholtz Centre for Infection Research in Braunschweig (Germany) together with the ifo Institute Munich [2].
https://www.ifo.de/en/node/55371
This study uses mathematical simulation methods to generate scenarios. The model does not assume a priori that there is a choice between public health measures and ‘the economy’. The questions is: What is the optimal public health policy profile (eg behavioural restrictions in response to the coronavirus) such that the number of death is minimised at the lowest possible decline of GDP. One of the possible scenarios is that no public health measures results in ‘maximum’ decline of GDP.
It is yet another example of joint work between natural scientists and economists.
It is not a cost-benefit analysis in contrast to the study by Whitton et al on the social cost of smoking.
With respect to [1], it seems to me this combination of university studies would be beyond the financial means of high income families, short of multimillionaires, in a system of privately funded higher education.
[1] Michael Meyer-Herman studied Physics, Mathematics, and Philosophy in Frankfurt/Main and Paris and accomplished his Ph. D. in Theoretical Elementary Particle Physics in Frankfurt/Main
[2] The ifo institute’s main objective is concerned with the role of government action in sustaining and increasing economic prosperity and social cohesion.
The interesting thing about that study is that it suggests stricter measures than the ones already in place despite some built in rightwingery. Ifo is after all the place run by Hans Werner Unsinn, infamous for sloppy work favouring right wing results.
Should add that the Ifo Institute has a good academic standing – im neither qualified nor interested enough to sort throug their peer reviewed output, which sure is not without right wing bias either. The real uggly stuff however were the non peer reviewed publications Unsinn often put out in the world under the IFO banner that just makes you scratch your head without any need for a particular qualification. Guess he is retired now?
hix, yes, Hans Walter (Un) Sinn retired from ifo in 2016. The head of ifo is now Clemence Fuest.