Tag Archives: Pandemic

Constant Elasticity Prevention

While many engaged Analysis #1 in my last post, only one engaged Analysis #2. So let me try again, this time with a graph.

This is about a simple model of prevention, one that assumes a constant elasticity (= power law) between harm and prevention effort. An elasticity of 1 means that 1% more effort cuts harm by 1%. For an elasticity of 2, then 1% more effort cuts harm by 2%, while for an elasticity of 0.5, 1% more effort cuts harm by 0.5%.

Such simple “reduced form” models are common in many fields, including economics. Yes of course the real situation is far more complex than this. Even so, reduced forms are typically decent approximations for at least small variations around a reference policy. As with all models, they are wrong, but can be useful.

Each line in the following graph shows how total loss, i.e., the sum of harm and prevention effort, varies with the fraction of that loss coming from prevention. The different lines are for different elasticities, and the big dots which match the color of their lines show the optimum choice on each line to min total loss. (The lines all intersect at prevention = 1/20, harm = 20.)

As you can see, for min total loss you want to be on a line with higher elasticity, where prevention effort is more effective at cutting harm. And the more effective is prevention effort, then the more effort you want to put in, which will result in a larger fraction of the total harm coming from prevention effort.

So if locks are very effective at preventing theft, you may well pay a lot more for locks on than you ever suffer on average in theft. And in the US today, the elasticity of crime with respect to spending on police is ~0.3, explaining why we suffer ~3x more losses from crime than we spend on police to prevent crime.

Recently, I asked a few polls on using lockdown duration as a way to prevent pandemic deaths. In these polls, I asked directly for estimates of elasticity, and in this poll, I asked for estimates of the ratio of prevention to health harm loss. And here I asked if if the ratio is above one.

In the above graph there is a red dot on the 0.5 elasticity line. In the polls, 56% estimate that our position will be somewhere to the right of the red dot on the graph, while 58% estimate that we will be somewhere above that grey 0.5% elasticity line (with less elasticity). Which means they expect us to do too much lockdown.

Fortunately, the loss at that red dot is “only” 26% higher than at the min of the grey line. So if this pandemic hurts the US by ~$4T, the median poll respondent expects “only” an extra $1T lost due to extra lockdown. Whew.

Added 26May: Follow-up surveys on US find (via lognormal fit) median effort to harm ratio of 3.6, median elasticity of 0.23. For optimum these should be equal – so far more than optimal lockdown!

Added 1Aug: Repeating same questions now gives median effort to harm ratio of 4.0, median elasticity of 0.18. That is, they see the situation as even worse than they saw it before.

Added 22Oct: Repeating the questions now gives median effort to harm ratio of 5.2, median elasticity of 0.10. The estimated deviation between these two key numbers has continued to increase over time.

GD Star Rating
loading...
Tagged as: , ,

2 Lockdown Cost-Benefit Analyses

Back on Mar. 21 I complained that I hadn’t seen any cost-benefit analyses of the lockdown policies that had just been applied where I live. Some have been posted since, but I’ve finally bothered to make my own. Here are two.

ANALYSIS #1: One the one side are costs of economic disruption. Let us estimate that a typical strong lockdown cuts ~1/3 of income of econ/social value gained per unit time. (It would be more due to harm from time needed to recover afterward, and to due to for stress and mental health harms.) If one adds 9 weeks of lockdown, perhaps on and off spread out over a longer period, that’s a total of 3 week’s income lost.

On the other side are losses due to infection. I estimate an average infection fatality rate (IFR) of 0.5%, and half as much additional harm to those who don’t die, due to other infection harms. (E.g., 3% have severe symptoms, and 40% of those get 20% disabled.) I estimate that eventually half would get infected, and assume the recovered are immune. Because most victims are old, the average number of life years lost seems to be about 12. But time discounting, quality-of-life adjustment, and the fact that they are poorer, sicker, and wouldn’t live as long as others their age, together arguably cuts that figure by 1/3. And a standard health-econ estimate is that a life-year is worth about twice annual income. Multiply these together and you get an expected loss of 3 week’s income..

As these equal the same amount, it seems a convenient reference point for analysis. Thus, if we believed these estimates, we should be indifferent between doing nothing and a policy of spending 9 added weeks of lockdown (beyond the perhaps 4-8 weeks that might happen without government rules) to prevent all deaths, perhaps because a vaccine would come by then. Or, if death rates would actually be double this estimate due to an overloaded medical system, we should be indifferent between doing nothing and spending 9 added weeks of lockdown to avoid that overloading. Or we should be indifferent between doing nothing and 4 added weeks of lockdown which somehow cuts the above estimated death rate in half.

Unfortunately, the usual “aspirational” estimate for a time till vaccine is far longer, or over 18 months. And a doubling of death rates seems a high estimate for medical system overload effects, perhaps valid sometimes but not usually. It seems hard to use that to argue for longer lockdown periods when medical systems are not nearly overwhelmed. Especially in places like the US with far more medical capacity.

During the 1918 flu epidemic, duration variations around the typical one month lockdown had no noticeable effect on overall deaths. In the US lately we’ve also so far seen no correlation between earlier lockdowns and deaths. And people consistently overestimate the value of medical treatment. Also, as death rates for patients on the oft-celebrated ventilators is 85%, they can’t cut deaths by more than 15%.

We’ve had about 6 weeks of lockdown so far where I live. A short added lockdown seems likely to just delay deaths by a few months, not to cut them much, while a long one seems likely to do more damage than could possibly be saved by cutting deaths.

Of course you don’t have to agree with my reference estimates above. But ask yourself how you’d change them, and what indifferences your new estimates imply. Yes, there are places in the world that seem to have done the right sort of lockdown early enough in the process to get big gains, at least so far. But if your place didn’t start that early nor is doing that right sort of lockdown, can you really expect similar benefits now?

ANALYSIS #2: Consider the related question: how much should we pay to prevent crime?

Assume a simple power-law (= constant elasticity) relation between the cost H of the harm resulting directly from the crimes committed, and the cost P of efforts to prevent crime:

H = k*Pa,  or  dln/ dlnP = –a ,

where a is the (positive) elasticity of harm H with respect to prevention P. To minimize total loss L = H + P, you set P = (k*a)1/(1+a), at which point we have a nice simple expression for the cost ratio, namely P/H = a.

So, when you do it right, the more effective is prevention at stopping harm, then the larger is the fraction of total loss due to prevention. If 1% more prevention effort cuts 1% of crime, you should lose about the same amounts from harm and prevention. If 1% more prevention cuts 2% of crime, then you should lose twice as much in prevention as you do in harm. And if it takes 2% more prevention effort to cut 1% of crime, you should lose about twice as much in harm as you do in prevention.

This model roughly fits two facts about US crime spending: the elasticity is less than one, and most loss comes from the crimes themselves, rather than prevention efforts. Typical estimates of elasticity are around 0.3 (ranging 0.1-0.7). US governments spend $280B a year on police, courts, and jails, and private security spends <$34B. Estimates of the total costs of crime range $690-3410B.

Now consider Covid19 prevention efforts. In this poll respondents said 3.44 to 1 that more harm will come from econ disruption than from direct health harms. And in this poll, 56% say that more than twice the loss will come from econ disruption. For that to be optimal in this constant elasticity model, a 10% increase in lockdown, say adding 12 days to a 4 month lockdown, must cut total eventual deaths (and other illness harm) by over 20%. That seems very hard to achieve, and in this poll 42% said they expect us to see too much econ disruption, while only 29% thought we’d see too little.

(More on Analysis #2 in the next post.)

In this post I’ve outlined two simple analyses of lockdown tradeoffs. Both suggest that we are at serious risk of doing too much lockdown.

10am: On reflection, I changed my estimate of the lockdown from 25% to 27% of income, and my estimate of non-death harm from as-much-as to half-as-much-as the death harm. So my reference added shutdown duration is now 4 months instead of 6.

12pm: Even if recovery gave immunity for only a limited period, then as long as you were considering lockdown durations less than that period, the above calculation still applies, but now it applies to each such period. For example, if immunity only lasts a year, then these are annual costs, not eventual costs. And that’s only if infection chances are independent each period. If, more likely, it is the same people who at more at risk each year, then in later years gains from lockdowns decline.

29Apr, 3am: We are now at 73 comments, and so far all of them are about analysis #1, and none about analysis #2. Also, tweet on #1 got 18 retweets, tweet on #2 got none.

29Apr, 1pm: In two more polls. over half estimate a 10% increase in lockdown duration gives <5% decrease in deaths, for both world and US. Instead of the >20% that would be required to justify allowing twice the damage of lockdowns as health harms. See also results on the cost of masks.

28May:  I’ve updated the numbers a bit.

22Oct: This analysis from March 22, based on happiness, also suggests far more harm from the economy dip than from deaths. And I confirm my analysis with more recent estimates here.

23Oct: I’ve just shown that the above condition that =dln/ dlnP = P/H holds for any function H(P).

GD Star Rating
loading...
Tagged as: , , , ,

Vouch For Pandemic Passports

Car pollution is an externality. Via pollution, the behavior of some hurts others, an effect that injurers may not take into account unless encouraged to by norm, contract, liability, or regulation. However, as the pollution from one vehicle mixes with that from many others, liability is poorly suited to discourage this; it is too hard to identify which cars hurt you, and there are too many of them. It seems to work better use regulations regarding car design and maintenance to limit the pollution emitted per mile driven, and to tax those miles driven.

Assault is also an externality; you can hurt someone by punching them. But in contrast to pollution, regulation is poorly suited to assault. We could require everyone to wear boxing globes and headgear, and we might ban insults and alcohol consumption, or perhaps even all socializing. But such regulations would go too far in restricting useful behavior. It works better to just hold liable those who punch others, via tort or criminal law. Yes, to discourage assault in this way, we must hold (or at least threaten to hold) expensive trials for each assault. But that still seems far cheaper than regulatory solutions.

These two examples illustrate a well-known tradeoff in choosing between (strict) liability and regulation. On the one hand, making people pay damages when they hurt others encourages them to take such harms into account, while also letting their behavior flexibly adapt to other context. On the other hand, regulation lets us avoid expensive court trials that require victims to prove who hurt who when where and how much. Though regulation induces more uniform behavior that is less well adapted to circumstances, it works acceptably well in many cases, like auto pollution, even if less well in others, like assault. (The mixed solution of negligence liability is discussed below.)

In our current pandemic, the main externality is infection, whereby one person exposes another to the virus. Conventional public health wisdom says to discourage infection via regulation: tell everyone when to get tested and isolated, and make them tell you who they met when and where. Tell them who can leave home when and for what reasons, and what they must wear out there. However, as we are all now experiencing first hand, not only are such changes to our usual behaviors quite expensive, such rules can also induce far from optimal behavior.

Recent pandemic rules have banned bike riding, but not cars or long walks. You can take only one exercise trip per day, but there’s no limit on how long. Members of a big home can all meet in their yard together, but members of two small adjacent homes may not meet in one of their yards. You can’t go meet distant friends, even if they only ever meet you. All parks are closed regardless of how densely people would be in there. The same rules were set in dense cities and in sparse rural areas. Alcohol store workers are deemed critical, even though alcohol can be mailed, but not auto repair, which cannot be mailed. Six feet is declared the safe distance, regardless of how long we stay near, if we wear masks, if we are outdoors, or which way the air is moving. Workplaces are closed regardless of the number of workers, how closely they interact, or how many other contacts each of them have. The same rules apply to all regardless of age or other illness. You may have to wear a mask, but it doesn’t have to be a good one.

Imagine that we instead used legal (strict) liability to make as many of us as possible expect to suffer personally and directly from infecting others, and to suffer more-so the worse their symptoms. In this scenario, such people would try to take all these factors and more into account in choosing their actions. For actions that risk infecting others, they would consider not only on how important such acts are to them, but also on how likely they are personally to be infected now, how vulnerable each other person they come near might be to suffering from an infection, how vigorously their activity moves the air near them, where such air currents are likely to go, how well different kinds of masks hinder infected air, and so on. If allowed, they might even choose variolation.

Of course, for the purpose of protecting ourselves from getting infected by others, we already have substantial incentives to attend to such factors. The problem is that simple regulations don’t give us good incentives to attend to these factors for the purpose of preventing us from infecting others. With regulations, we have incentives to follow the letter of the law, but not its spirit. So we don’t do enough in some ways, and yet do too much in others. But if liability could make us care about infecting others as well as ourselves, then it might simultaneously reduce both infections and the economic and social disruptions caused by lockdowns. With strong and clear enough liability incentives, we wouldn’t need regulations; we could just let people choose when and how to work, shop, travel, etc.

But is it feasible to use liability to discourage infections? Yes, if we can satisfy two conditions: (1) most people are actually able to pay for damages if they are successfully sued for infecting others, and (2) enough of those who infect others are actually and successfully sued, and so made to pay.

On the first condition, ensuring that people can pay damages if they are found guilty, it is sufficient to require people who mix with others to buy infection liability insurance, similar to how we now require car drivers to get accident liability insurance. That is, to get a “pandemic passport” to excuse you from a strong lockdown, you must get an insurance company to guarantee that you will pay damages if you are shown to have infected someone. In a sense they “vouch” for you, and so are your “voucher”. The more types of voucher-client contract terms we are willing to enforce, the more levers vouchers gain to reduce risks.

The premiums for such insurance will be low if you can convince a voucher that you have already recovered from the virus, and so are relatively immune, or that you will leave your lockdown only rarely, to safe destinations. Otherwise, a voucher may require you to install an app on your phone to track your movements, or they may spot check your claims that you have sufficient supply of good masks that you use reliably when you leave home.

Okay, but what about the second condition, that enough infectors are actually made to pay? For this we need enough data to be collected on both sides, the infector and the infected, so that one can frequently enough match the two, to conclude that this person likely infected that one at this location at this time.

Now, we don’t need to be always absolutely sure of who infected who. In ordinary civil trials, the standard is a “preponderance of the evidence”; courts need only be 51% or more sure to convict the defendant. And sometimes we add on extra “punitive” damages, up to four ties as large as basic damages, often to compensate for a lower chance of catching offenders. So if we can find evidence to convince a court at the 51% or better standard for only one fifth of offenders, but we can add four times punitive damages, then offenders who do not know if they will be caught still expect to on average to pay near the basic damage amount.

Okay, but we still need to collect enough info to see who infected who at least one fifth of the time. Is this feasible? Well it is clearly quite feasible early in a pandemic, when few have been infected. Early on, if the times and places, i.e., space-time path, consistent with you being infected then and there overlap with the space-time path when someone else was likely infectious, then it was most likely their fault. This is the “contact trace” process usually recommended by public health workers early in a pandemic.

The problem gets harder later in a pandemic, when your infected path may overlap with the infectious paths of many others. Here it might be possible to use info on which virus strain you and they had to narrow the field. But even so there may still be several consistent candidates. In this case it seems reasonable to divide the liability over all of them, perhaps in proportion to the size of the path overlap. For the purpose of creating incentives to avoid infecting others, it isn’t that important to know later who exactly infected who when.

But yes, we still need info on who was infected and infectious where and when, perhaps supplemented by data on who had what virus strains. How can we get this info? People who might get infected have incentives to collect info on their path, to help them sue if infected. But people who might infect others would seem to want to erase such info, to keep them from being sued. I’ve recently outlined a more general approach to induce the collection of info sufficiently likely to be useful in later lawsuits. But for this essay, I’ll just propose that collecting key info be another condition required to get a pandemic passport, with violations punished by fines also guaranteed by your voucher.

Let me also note that yes, legal liability doesn’t work to discourage harms if typical harms get so small that people wouldn’t bother to sue to recover damages. In this case we could use a random lottery approach to dramatically lower the average cost of suing.

So let’s put this all together. You must stay at home, locked down, unless you get a “pandemic passport”, in which case you can go where you want when, to meet anyone. To get such a passport, you must get someone to vouch for you. They guarantee that you will pay should someone successfully sue you for infecting them, if you agree to their terms of premiums, behavior, monitoring, punishment, and co-liability. Defendants who pay damages may have to pay extra, to compensate for most infectors not getting caught in this way. When several infector candidates are consistent with the data, they can divide the damages. And for low damage levels, a random lottery approach can lower court costs.

To get and keep a passport, your voucher also guarantees that you will collect info that can help others to show that you infected them, but which can also help you to sue others if they infect you, and win you bounties via showing that others did not collect required info. For example, perhaps you must track your movements in space and time, regularly record some symptoms like body temperature, and also save regular spit samples. This info is available to be subpoenaed by those who can show sufficient reason to suspect that you infected them. Such info seems sufficient to catch enough infectors.

And by catching a sufficient fraction of infectors who then actually pay on average for the harms that they cause by infecting, (strict) legal liability can give sufficient incentives to individuals to avoid infecting others. If so, we don’t need crude lockdown regulations telling people what to do when and how; individuals can instead more flexibly adapt to details of their context in deciding when and where to work, shop, travel etc. Yes, voucher rules would not let them do such things as freely as they would in the absence of a pandemic. But behavior would be more free and impose lower economic costs than under crude regulations which similarly suppress the pandemic spread.

Note that today the most common form of legal liability is actually negligence, which we can see as a mixed form between simple regulation and simple strict liability. With negligence, the court judges if your behavior has been consistent with good behavior standards, which are essentially behavior regulations. But you are only punished for violating these regulations in situations where your behavior contributed to the harm of a particular person. Today courts tend to limit strict liability to cases where courts find it hard to define or observe good behavior details, such as using explosives, keeping a pet tiger, or making complex product design choices. As courts find it harder to define and observe good behavior in a new pandemic, strict liability seems better suited to this case.

Note also that none of this requires employers to be liable for their infected employees. Someone who is sued for infecting others may turn around and blame their employer for pushing them into situations that cause them to infect others. Employer-employee contract could usefully address such issues.

GD Star Rating
loading...
Tagged as: , , ,

Subpoena Futures

A subpoena … is a writ issued by a government agency, most often a court, to compel testimony by a witness or production of evidence under a penalty for failure. … party being subpoenaed has the right to object to the issuance of the subpoena, if it is for an improper purpose, such as subpoenaing records that have no relevance to the proceedings, or subpoenaing persons who would have no evidence to present, or subpoenaing records or testimony that is confidential or privileged. (More)

Parties may obtain discovery regarding any nonprivileged matter that is relevant to any party’s claim or defense and proportional to the needs of the case, considering the importance of the issues at stake in the action, the amount in controversy, the parties’ relative access to relevant information, the parties’ resources, the importance of the discovery in resolving the issues, and whether the burden or expense of the proposed discovery outweighs its likely benefit. (More)

Exceptions are quite limited: self-incrimination, illegally-obtained info, and privileges of spouses, priests, doctors, diplomats, and lawyers. The remarkable fact is that the law has little general respect for privacy. Unless you can invoke one of these specific privileges, you must publicly report to the court any info that it thinks sufficiently relevant to a current court case. You simply have no general right to or expectation of privacy re stuff a court wants to know. Courts don’t even compensate you for your costs to collect evidence or appear to testify.

And yet. Consider what I wrote March 5:

The straightforward legal remedy for [pandemic] externalities is to let people sue others for infecting them. In the past this remedy has seemed inadequate for two reasons:

1. It has often been expensive and hard to learn and prove who infected who, and
2. … most folks just can’t pay large legal debts.
The vouching system directly solves (2), … And the key to (1) is ensuring that the right info is collected and saved.

First, consider some new rules that would limit people’s freedoms in some ways. Imagine people were required to keep an RFID tag (or visible QR code) on their person when not at home, and also to save a sample of their spit or skin once a week? Then phones could remember a history of the tags of people near that phone, and lawsuits could subpoena to get surveillance records of possible infection events, and to see if spit/skin samples on nearby dates contain a particular pathogen, and its genetic code if present. We might also adopt a gambled lawsuit system to make it easier to sue for small harms. (More)

Here, to help law deal with pandemics, I was tempted to propose specific rules re info that people must collect and preserve. Yet if courts can get any info they think relevant, why is there ever a problem with courts lacking info to deal with key harms, such as pandemic infection?

The answer is that current law allows a huge exception to its subpoena power. Courts can force you to reveal info that you have already collected, on paper, a computer, in your head, or in your physical objects. But you usually have no obligation to collect and save info now that the court might want later. As a result, many people and orgs go out of their way to not save incriminating info. For example, firms do key discussions verbally, not recorded, rather than via email. Thus you have no obligation to save spit samples or detailed records of where your phone goes, to help with future pandemic infection lawsuits.

This seems a huge and inconsistent loophole. I could understand if the law wanted to respect a more general right to privacy. Then the court might weigh the value of some info in helping court cases against the social harm from forcing its publication via a subpoena. As a result, it might sometimes block a subpoena even when the info collected would be relevant to a court case.

But I can’t see a reason to eagerly insist on access to info that seems relevant to a court case, and yet put no effort into inducing people to collect and preserve such info beforehand. So I propose that we create a legal process by which legal judgements are made on, if collected and saved, how likely some info would be to be subpoenaed, and how valuable it would be in that case.

When info would be valuable enough if collected and saved, then the court should require this. I don’t have a strong opinion on who exactly should bring a suit asking that such info be saved, or who should represent the many who would have to save that info. But one obvious system that occurs to me is to just have courts usually make ex post estimates of info value by the end of each court case, and then use “subpoena futures” prediction markets to make an estimate of that value ahead of time. (And make it legal and cheap to start such markets.)

So, if a subpoena futures market on a type of info estimates its expected court value to be above a standard threshold, then by law that info must be collected and saved. These prediction markets needn’t be huge in number, if they could estimate the average value of such info collect over a large group, which would then justify requiring that entire group collect the info. Such as everyone in an area who might infect others with a pandemic. If some subgroup wanted to claim that such info wasn’t less valuable regarding them, and so they should be excused, why they’d have to create different prediction markets to justify their different estimates.

For example, when a pandemic appears, if those who might infect others are likely vouched, then those who might be infected would want to require that first group to collect and save info that could be used later to prove who infected who. So they’d create prediction markets on the likely court value of spit samples and phone location records, and use market estimates to get courts to require the collection of that info.

Compared to my prior suggestion of just having the law directly require that such info be collected, this subpoena futures approach seems more flexible and general. What other harms that we do each other could be better addressed by lawsuits if we could require that relevant info be collected and saved?

(Btw, courts need not estimate info value in money terms. They might instead express the value of each piece of info in terms of its multiple of a “min info unit”, i.e., the value of info where they’d be just on the border of allowing it to be subpoenaed for a particular case.)

Added 7a: As mentioned in this comment, we now have this related legal concept:

Spoliation of evidence is the intentional, reckless, or negligent withholding, hiding, altering, fabricating, or destroying of evidence relevant to a legal proceeding …The spoliation inference is a negative evidentiary inference that a finder of fact can draw from a party’s destruction of a document or thing that is relevant to an ongoing or reasonably foreseeable civil or criminal proceeding.

My proposal can be seen as expanding this concept to allow a much weaker standard of “foreseeable”. And instead of allowing a presumption at trial, we just require the evidence to actually be collected.

GD Star Rating
loading...
Tagged as: , , , ,

A Perfect Storm of Inflexibility

Most biological species specialize for particular ecological niches. But some species are generalists, “specializing” in doing acceptably well in a wider range of niches, and thus also in rapidly changing niches. Generalist species tend to be more successful at generating descendant species. Humans are such a generalist species, in part via our unusual intelligence.

Today, firms in rapidly changing environments focus more on generality and flexibility. For example, CEO Andy Grove focused on making Intel flexible:

In Only the Paranoid Survive, Grove reveals his strategy for measuring the nightmare moment every leader dreads–when massive change occurs and a company must, virtually overnight, adapt or fall by the wayside–in a new way.

A focus on flexibility is part of why tech firms tend more often to colonize other industries today, rather than vice versa.

War is an environment that especially rewards generality and flexibility. “No plan survives contact with the enemy,” they say. Militaries often lose by preparing too well for the last war, and not adapting flexibly enough to new context. We usually pay extra for military equipment that can function in a wider range of environments, and train soldiers for a wider range of scenarios than we train most workers.

Centralized control has many costs, but one of its benefits is that it promotes rapid thoughtful coordination. Which is why most wars are run from a center.

Familiar social institutions tend to be run by those who have run parts of them well recently. As a result, long periods of peace and stability tend to promote specialists, who have learned well how to win within a relatively narrow range of situations. And those people tend to change our rules and habits to suit themselves.

Thus rule and habit changes tend to improve performance for rulers and their allies within the usual situations, often at the expense of flexibility for a wider range of situations. As a result, long periods of peace and stability tend to produce fragility, making us more vulnerable to big sudden changes. This is in part why software rots, and why institutions rot as well. (Generality is also often just more expensive.)

Through most of the farming era, war was the main driver pushing generality and flexibility. Societies that became too specialized and fragile lost the next big war, and were replaced by more flexible competitors. Revolutions and pandemics also contributed.

As the West has been peaceful and stable for a long time now, alas we must expect that our institutions and culture have been becoming more fragile, and more vulnerable to big unexpected crises. Such as this current pandemic. And in fact the East, which has been adapting to a lot more changes over the last few decades, including similar pandemics, has been more flexible, and is doing better. Being more authoritarian and communitarian also helps, as it tends to help in war-like times.

In addition to these two considerations, longer peace/stability and more democracy, we have two more reasons to expect problems with inflexibility in this crisis. The first is that medical experts tend to think less generally. To put it bluntly, most are bad at abstraction. I first noticed this when I was a RWJF social science health policy scholar, and under an exchange program I went to the RWJF medical science health policy scholar conference.

Biomed scholars are amazing in managing enormous masses of details, and bringing up just the right examples for any one situation. But most find it hard to think about probabilities, cost-benefit tradeoffs, etc. In my standard talk on my book Age of Em, I show this graph of the main academic fields, highlighting the fields I’ve studied:

Academia is a ring of fields where all the abstract ones are on one side, far from the detail-oriented biomed fields on the other side. (I’m good at and love abstractions, but have have limited tolerance or ability for mastering masses of details.) So to the extent pandemic policy is driven by biomed academics, don’t expect it to be very flexible or abstractly reasoned. And my personal observation is that, of the people I’ve seen who have had insightful things to say recently about this pandemic, most are relatively flexible and abstract polymaths and generalists, not lost-in-the-weeds biomed experts.

The other reason to expect a problem with flexibility in responding to this pandemic is: many of the most interesting solutions seem blocked by ethics-driven medical regulations. As communities have strong needs to share ethical norms, and most people aren’t very good at abstraction, ethical norms tend to be expressed relatively concretely. Which makes it hard to change them when circumstances change rapidly. Furthermore we actually tend to punish the exceptional people who reason more abstractly about ethics, as we don’t trust them to have the right feelings.

Now humans do seem to have a special wartime ethics, which is more abstract and flexible. But we are quite reluctant to invoke that without war, even if millions seem likely to die in a pandemic. If billions seemed likely to die, maybe we would. We instead seem inclined to invoke the familiar medical ethics norm of “pay any cost to save lives”, which has pushed us into apparently endless and terribly expensive lockdowns, which may well end up doing more damage than the virus. And which may not actually prevent most from getting infected, leading to a near worst possible outcome. In which we would pay a terrible cost for our med ethics inflexibility.

When a sudden crisis appears, I suspect that generalists tend to know that this is a potential time for them to shine, and many of them put much effort into seeing if they can win respect by using their generality to help. But I expect that the usual rulers and experts, who have specialized in the usual ways of doing things, are well aware of this possibility, and try all the harder to close ranks, shutting out generalists. And much of the public seems inclined to support them. In the last few weeks, I’ve heard far more people say “don’t speak on pandemic policy this unless you have a biomed Ph.D”, than I’ve ever in my lifetime heard people say “don’t speak on econ policy without an econ Ph.D.” (And the study of pandemics is obviously a combination of medical and social science topics; social scientists have much relevant expertise.)

The most likely scenario is that we will muddle through without actually learning to be more flexible and reason more generally; the usual experts and rulers will maintain control, and insist on all the usual rules and habits, even if they don’t work well in this situation. There are enough other things and people to blame that our inflexibility won’t get the blame it should.

But there are some more extreme scenarios here where things get very bad, and then some people somewhere are seen to win by thinking and acting more generally and flexibly. In those scenarios, maybe we do learn some key lessons, and maybe some polymath generalists do gain some well-deserved glory. Scenarios where this perfect storm of inflexibility washes away some of our long-ossified systems. A dark cloud’s silver lining.

GD Star Rating
loading...
Tagged as: , , , ,

Covid19 Pizza-Risk Estimator

To keep us from catching and spreading Covid19, most of us are on “lockdown”, limiting non-home contact. But we aren’t completely isolated; we all do a few things that risk outside contact. And households really are the better unit of risk here; if one of you gets sick, the rest face a much higher risk.

To help households estimate and manage risk, I’ve made the following table, listing risks for 19 activities, all relative to the first: accepting delivery of and eating a pizza, paid for online. These risk estimates came from ~1000 respondents to each of 18 Twitter polls. (Technically, these estimates are medians of lognormal distributions fitted to poll response frequencies.) Yes, it would be better to get expert estimates, but I don’t have experts to poll. I hope experts who see these will publicly improve on them. Until they tell us more, we must act on what we know.

The above is actually part of a screenshot from this spreadsheet (copy it to edit it) that I’ve made to help you estimate household risk. (Anyone know how to embed it here, so each reader can edit their own version here?) On this sheet, you can combine these risk estimates with estimates of how often per week your household does each activity, and also any corrections for how you do it differently, to get your total household weekly “Pizza Risk”. That is, how many weekly risks you take equivalent to a pizza delivered & eaten.

In the spreadsheet, each row lists a risky activity, grouped into types. To use the sheet, consider the activity in each row and think of similar activities you do that risk outside contact. For each such activity, find the closest activity in the table, and for that row, enter how many times per week your household does that activity in the “Count” column. And if your activity seems to have a different risk from other households, such as because you do it for more or less time, or because it involves fewer or more outsiders, then enter a number other than 1 in the “Factor” column. For example, if doing it your way has twice the risk, enter the number 2.

If you mange to use this spreadsheet to get a Pizza Risk estimate, please complete the following two polls so we can learn about how Pizza Risk varies across households.

FYI, this post was up just 33 hours after I first tweeted the idea for this project.

Added 14Apr: Commentor Roman Kuksin did an explicit analysis of many of these risks, and finds a 0.74 correlation with the above risk estimates.

Added 18Apr: Here is another set of activity risk estimates: HT @diviacaroline.

GD Star Rating
loading...
Tagged as: ,

Peter Doherty on Variolation

Two noteworthy media mentions of variolation:

1. The New Yorker features Douglas Perednia talking about controlled infection and variolation, as its Exhibit A on on why conservative media shouldn’t presume to write on health/medicine, as they will only say stupid obviously terrible things:

After the Federalist tweeted it out, Twitter, which has been cracking down on coronavirus misinformation, temporarily locked the Federalist’s account. … He’d submitted it to a number of medical journals and blogs. “They all turned it down with no comment,” … tried the Federalist, almost at random. … The site accepted his article the next day, no questions asked. …

On the site … most commenters found Perednia’s idea absurd, dangerous, hilarious, or all three. … many angry e-mails and calls … Andrew Lover, an assistant professor of epidemiology at the University of Massachusetts-Amherst, told the Times that Perednia’s article was “exceedingly ill-advised and not evidence-based in any way shape or form.” …

Perednia [said] the way to adapt his idea to this reality was to make sure that the infecting was done with ‘the lowest possible dose’ … a concept known as variolation—which, he thought, would cut the death rate among those who chose to take part. (more)

2. A month ago, Adam Ford interviewed me on voluntary infection. Yesterday, Ford posted his interview with Nobel laureate Peter Doherty, author of Pandemics (2012), wherein Ford asked Doherty about variolation. Here are selected quotes from that discussion (fuller quotes below the fold):

47:10 Ford: “Controversially, in leu of actual vaccine that could come, hopefully in 9 months, but maybe even in 18 months if things go okay, if social isolation doesn’t work well enough too, would something like strategic or voluntary small dose low dose infection, like variolation, work in order to gain immunity, or nudge herd immunity? Is that something that we should be considering?

Doherty: (laughing) “Well let’s tell people what variolation was. … What they did was do this in young children, young children had a good immune response, generally survived smallpox, so what they were doing essentially is giving them smallpox, and they survived, whereas if they got it when they were older, they’d have a much worse disease. … So its not an unthinkable thing. …

52:40 With Covid19 I don’t know, but it would take a brave soul to be a test candidate, With younger people who are not severely affected, it’s possible. But you’d have to be enormously careful that they didn’t get any dose through their nose. But there would be ways of doing this. …

53:40 Ford: Is this something that could be achieved in the near term, if the vaccine timeline ends up looking like its going to be longer?

Doherty: If it was an absolutely catastrophic situation, if it was like the situation that is depicted in Contagion, where everyone who is within 100 feet of the virus gets it and dies, yes it could be reasonable. But I think for a virus where 80+% of people are definitely mildly infected at worse, or not sick enough to go into hospital, I don’t think you would take risk of that. The thing about a vaccine is that you have to give it to large numbers of normal people. You can’t take risks with vaccines.

You can take risks with end stage therapy. If someone is very very sick, and you’ve got something you think might work, you can try it pretty easily. People will approve that, … But you can’t take risks with vaccines. And the magnitude of the severity of this threat is not great enough to do that. You could say, … we’ll take a vaccine that looks a bit risky, maybe, and we’ll give it to the elderly. These are the people who are at risk, they can try it. … People like me, say would volunteer, I certainly would. I’d give it a go, and see if that works. But I wouldn’t want to be giving a vaccine that had any risk at all to younger people. You know, these are all theoretical arguments. But there is no way anything is ever given to anybody in this sense without going through extremely thorough review processes. … I think it is pretty unlikely.

So Doherty accepts “variolation” as a term that applies outside the context of smallpox. He thinks it could work, but oddly seems to see the main concept as infecting the young, rather than controlling dose, delivery vector, or strain. And he sees it only as justified in extreme circumstances, which Covid19 will never be, as it isn’t deadly enough. Even if the Great Suppression crashes the economy worse than the Great Depression, and even if millions will likely die from accidental infections, in his eyes and those of regulators that’s no excuse for letting healthy people voluntarily take substantial personal risks. Continue reading "Peter Doherty on Variolation" »

GD Star Rating
loading...
Tagged as: ,

Variolation Test Design

Okay, what the variolation concept needs most now is a trial/test/experiment ASAP. So to help get the ball rolling, let me sketch a tentative plan. I’m NOT saying this plan is now good enough. I’m saying let’s talk together about how to make it better. (Not so interested here in those ever popular “this can never work” comments.)

As with most projects, the obvious first top issue is staffing, especially leaders. This needs leaders who not only have the ability and expertise to execute it, but who can also inspire confidence in its other staff, subjects, patrons, sponsors, and audiences. (The most I’ve ever led is an assistant, so alas I don’t seem a good candidate.) The main point here is to inspire audiences to action, and that won’t happen if audiences don’t believe the project’s purported results, nor if they find its people too odious to associate with.

So the main purpose of this post is to try to attract participants, especially leaders, to pick up this ball and run with it. I’ll run with you, but I can’t run it by myself. When someone makes a good suggestion, such as in the comments, I’m likely to edit this post to include it. You are warned. Continue reading "Variolation Test Design" »

GD Star Rating
loading...
Tagged as: ,

Reply to Cowen On Variolation

In the last nine days I’ve done two online debates on variolation, with Zvi Mowshowitz and Gregory Cochran. In both cases my debate partners seemed to basically agree with me; disagreements were minor. Last night Tyler Cowen posted 1000+ words on “Why I do not favor variolation for Covid-19”. Yet oddly he also doesn’t seem to disagree with my main claims that (1) we are likely to need a Plan B for all-too-likely scenario where most of the world seems likely to get infected soon, and (2) variolation is simple, mechanically feasible, and could cut Covid19 Deaths by a factor of 3-30.

Tyler lists 8 points, but really makes 11. If he had one strong argument, he’d have focused on that, and then so could I in my response. Alas, this way I can’t respond except at a similar length; you are warned. Continue reading "Reply to Cowen On Variolation" »

GD Star Rating
loading...
Tagged as: ,

Beware R0 Variance

The big push now re Covid19 is to use “social distancing” to cut “R0”, the rate at which infection spreads. More precisely, R0 is the average number of other people that one infected person would infect, if they were not already infected. With no efforts to reduce it, estimates for natural R0 range from 2 to 15, with a best estimate perhaps around 4. The big goal is to get this number below 1, so that the pandemic is “suppressed” and goes away, and stays away, until a vaccine or other strong treatment, allowing most to escape infection. In contrast, if R0 stays above 1 we might “flatten the curve”, so that each infected person can get more medical resources when they are sick, but soon most everyone gets infected.

Apparently even with current “lockdown” efforts, all of 11 European nations studied now still have best estimate R0 over 2, with a median ~3.7. So they must do a lot more if they are to suppress. But how much more? My message in this post is that it is far from enough to push median R0 down below 1; one must also push down its variance.

Imagine a population composed of different relatively-isolated subpopulations, each with a different value of R0. Assume that few are infected, so that subpopulation pandemic growth rates are basically just R0. Assume also that these different R0 are distributed log-normally, i.e., the logarithm of R0 has a Gaussian distribution across subpopulations. This is (correctly) the usual distribution assumption for parameters bounded by zero below, as usually many small factors multiply together to set such parameters. The total effective R0 for the whole population is then found simply by integrating (via a lognormal) the effective growth over R0 subpopulations.

For example, assume that the R0 lognormal distribution has log mean (mu) -2 and sigma 1. Here the mode of the distribution, i.e., the most common R0 number, is 0.05, the median R0 is 0.14, only 5% of subpopulations have R0 above 0.70, and only 2% have R0 >1. Even so, if each of these subpopulations maintain their differing R0 over ten infection iterations, the mean growth factor R0 of the whole population is 20 per iteration!

As another example (for log mean -1, sigma 0.5), the R0 mode is 0.29, the median is 0.37, only 5% of subpopulations have an R0 over 0.85, only 2% have R0>1. Yet over ten infection iterations maintaining these same R0 factors per subpopulation, the mean growth factor R0 of the whole population is 1.28 per iteration. That is, the pandemic grows.

Of course these growth numbers eventually don’t apply to finite subpopulations, once most everyone in them gets infected. Because when most of a population is infected, then R0 no longer sets pandemic growth rates. And if these subpopulations were completely isolated from each other, then all of the subpopulations with R0<1 would succeed in suppressing. However, with even a modest amount of interaction among these populations, the highly infected ones will infect the rest.

The following graph tells a somewhat more general story. On the x-axis I vary the median value of R0 among the subpopulations, which sets the log-mean. For each such value, I searched for the log-sigma of the lognormal R0 distribution that makes the total average R0 for the whole population (over ten iterations) exactly equal to 1, so that the pandemic neither grows nor shrinks. Then on the graph I show the standard deviation, in R0 terms, that this requires, and the fraction of subpopulations that grow via R0>1.

As you can see, we consistently need an R0 standard deviation less than 0.21, and the lower the median R0, the lower a fraction of subpopulations with a positive R0 we can tolerate.

So, as long as there is substantial mixing in the world, or within a nation, it is far from enough to get the R0 for the median subpopulation below 1. You also need to greatly reduce the variation, especially the fraction of subpopulations in which the pandemic grows via R0>1. For example, when the median R0 is 0.5, you can tolerate less than 3% of subpopulations having an R0>1, just to hold the pandemic at a constant overall level. And to suppress in limited time, you need to go a lot further.

Different subpopulations with differing R0 seems plausible not just because our world has different nations, classes, cultures, professions, industries, etc., but because Covid19 policy has mostly been made at relatively local levels, varying greatly even within nations. In addition, most things that seem log-normally distributed actually have thicker than-lognormal tails, which makes this whole problem worse.

All of which is to say that suppressing a pandemic like this, with high R0 and many asymptomatic infected, after it has escaped its initial size and region, is very hard. Which is also to say, we probably won’t succeed. Which is to say: we need to set up a Plan B, such as variolation.

Spreadsheet for all this here.

GD Star Rating
loading...
Tagged as: ,