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.
It's not obvious that people can account for low-probability-but-large costs when they pick jobs. You could imagine people continuning to enter such fields and 10% of them going bankrupt every year due to liability, but most entrants not noticing.Of course, that's an empirical question. My theory would predict that lots of truckers go bankrupt due to liability. If we don't observe that then my concern evaporates in a puff of smoke.
I do not see RH's supply and demand argument here. But I won't distract him by making him explain. I'll go buy a book.
You realize that employees have no control over the employment contract, right. We sign what they tell us to sign or we starve. Say I’ve been appliying to jobs for months and run out of savings, and the employer sends me some contract that says “if you infect someone in the line of duty it’s on you”. Say I also happen to be a superforecaster and I realize the contract is negative expected utility. I still have to sign it, so I can keep a home. And if I’m sued I just have to put all my assets in bitcoin. Those are your options when your poor. Contracts and covid liability don’t affect our incentives in practice.
Fair enough about the distinguishing. But I think one problem when talking about such proposals is that it isn't clear just how much you are packing into the assumption that people will adopt the proposal.
I think the default assumption is to imagine that one gets to stipulate how the government behaves but then any uptake by the population at large or by corporations etc.. isn't taken for granted. This was the assumption I was approaching it from. This corresponds to a kind of assumption that the discourse you are engaged in is in some sense about illuminating potential choices for policy makers and attempting to shift what they see as reasonable.
However, I'll grant that's not the only assumption you can approach the question from but I do think it's what many people assume you to be engaged in absent other indications.
Incidentally, I have started to see companies develop and publish procedures for liability management as they plan on reopening in coming weeks, including procedures for daily attestation from each employee before arriving at work that a) they are not expressing symptoms of the virus and b) have not had contact with anyone expressing symptoms or currently diagnosed in last 14 days.
Always distinguish between if people will adopt a proposal from whether it will work if adopted. Proposers are usually making claims about that second topic, not the first.
Issues of how to make legal decisions are usually decided in the context of concrete decisions. So we can't wait until they work it out until we allow cases; we have to start the cases so they will work it out.
If this insurance is required to get a passport, then customers would be eager to buy, and so insurance firms would quickly offer products. Within a month I'd say.
If people can't afford infection insurance, that's a pretty good clue that in fact we shouldn't let them out of lockdown.
If people are doing good things but getting insurance is a cost of doing those good things, well then of course we should be willing to pay those costs, to get those good things.
The main gain we will get form this pandemic is getting better prepared for a worst one. So we should go down this path, even if we aren't sure just when we can get gains from it.
You don't really believe this would work in the current situation in the real world do you? I mean you've explained in other posts all sorts of factors that incline people to dislike this kind of system (we can't even get sufficent support to get working, liquid/uncapped prediction markets running during normal times).
With this proposal there would be huge difficulties.1) First, lawmakers would have to create rules (likely more complex than current regulations) for assigning liability (do you pay for downstream infections? This gets really complex. For instance, what fraction of downstream damage awards are you responsible for?
As a matter of pure harm any damages assessed against a downstream infection are a harm you imposed on them but that system wouldn't work for insurers (bankcruptcy lotto if you happen to insure patient 0).
What kinds of evidentiary standards will be used?
1') Do you asses damages as they do for wrongful death (in terms of lost future wages), purely as a funciton of lost QALYs plus direct economic harms)? Any system that awards higher damages to higher earners likely creates a big backlash.
2) The courts, currently unable to adequately handle their normal proceedings, would need to rapidly expand to handle the vast flood of lawsuits. (Waiting will not on spoil evidence but impose massive legal interpratational uncertainty on insurers).
3) Insurers would need to somehow read this whole new body of federal tort law, model their exposure to various legal interpratational risks (passing costs on to consumers), generate acturial models with very little relevant data, explain and market such insurance to individuals (creating the systems to store, analyze and safeguard the data needed to adequately measure risk).
3') The insurance fraud problem becomes is likely to be pretty bad as people who get sick from uncovered or unidentifiable sources see themselves as being unfairly screwed over. Unlike arson or a car crash it's unlikely there will be any evidence besides testimony of those involved. A sizeable fraction of people infected from uncovered/unidentified sources likely find covered colluder to commit insurance fraud with (they will morally justify it to themselves by saying they don't *know* that it wasn't from them even if it's unlikely).
4) We don't actually want to disincentivize people from treating the sick, delivering food or working in food kitchens. You might argue that in long term the huge insurance premiums a doctor would have to pay if they treated COVID patients but that would be an insanely large shock that (in the face of wage and price stickiness) would produce huge economic harms.
Ohh wait, that means you'd need to completely overhaul the medicaid and insurance price structures to cover massie premium that doctors would have to pay for insurance.
4') Lots of people simply won't have the money for insurance and since we don't have an extensive and reliable system that lets people take on non-dischargeable debt those people end up staying home even if they could have been net economic contributors.
—
So to sum up you need a massive overhaul of tort law, health insurance and gov health programs, spin up of an unprecedented type of insurance market, likely huge negative externalities in the resulting economic disruption as members of valuable professions like doctors get hit with huge premiums and wages/payments lag behind, likely a massive expenditure of money in both the court system and in building the new insurance system. And I'm sure I'm forgetting some.
And what do you get for all this? Insurers don't yet have the know how or ability to collect the data that would allow them to price products in a way that substantially improves efficiency (giving them that data creates another massive externality from data privacy risks). Far from a super rational system you get a bunch of fudges and guesses about good ways to adjudicate COVID liability in court.
It's a neat thought experiment but it's not even close to practically workable in this pandemic.
Your article states, "the more types of voucher-client contract terms we are willing to enforce, the more levers vouchers gain to reduce risks." For the model you propose to be effective in the context of spreading a respiratory virus vs a car accident, I'm theorizing that - in comparison to auto insurance - a) the terms must require more overhead (tangible and intangible costs) to adequately defend your level of liability; b) the variables affecting risk are not distributed evenly across classes.
Regarding that overhead, insuring against car accidents is a pretty passive policy to carry (which incidentally, arguably does have problems in pricing across classes, if not qualifying). Your suggestions for terms which could effectively insure against spreading contagion include logging movement, maintaining and logging surplus PPE supplies, logging contact points, etc. This requires ongoing activity and spend, the lack of which substantially undermine one's protection when the critical challenge comes. Those able to afford the tools and services for automating these tasks are very advantaged. Maybe the insurer could provide the means to do this though, much like Progressive Snapshot. As an unrelated aside, all of this info could be subpoenaed, so let's at least acknowledge that there's some attention required there at a later date in re adoption and compliance.
Your Pizza Risk Estimator implicates some professions over others for increased risk. Those professions with higher risks are often those at the lower income brackets, while knowledge workers able to work in isolation are advantaged. This has cost or qualification implications across classes, no?
Then the basic context here is that we're implicating not the activity of speedy transport on public roads, but the daily, required activities of basic living - socializing, walking about, buying groceries, assembling. So there's a huge incentive to dumb the barriers down as much as possible if you want compliant participation, undermining the whole point. It works for automobiles but maybe not best solution for protecting against disease.
I understand you're arguing this as an alternative to costly lockdowns, and anything probably makes more sense as an alternative to perpetual lockdowns. But there is a whole legal issue with this idea that perpetual lockdown is even a threat. The right to assemble at some point must take priority over emergency powers, whose granting must by definition be exceptional (right??).
You may counter that one of these regime models may be the required condition to lift the emergency powers. Ignoring all legal considerations, I think just from a practical standpoint (given the above reasons) a license and regulatory regime would be preferable.
In reality, you have the stronger case for an insurance and liability model considering SCOTUS has precedent authorizing the compulsory purchase of insurance as a condition for existing in the US, where they have no such thing for licensing. ;)
Auto liability insurance is not watered down to let anyone qualify, and that's the process I propose to use here. I don't see the problem of applying this "across" classes.
Fun thought exercise but you know in the US this would unquestionably be disproportionately applied across social/racial classes (reasonably or not). Imagine if you "badged" the inverse - those who are insufficient. Same system, but not very good optics. To head off your "all-or-nothing" challenge, I would note that these systems of certification and attestation work to a degree, but you'll note that the driver's license test is watered down sufficiently to allow any idiot to pass by design. How would your proposal satisfy the need for inclusivity and maintain efficacy? I roll my eyes at the very question, but in world we have , a proposal not considering this angle can be considered DOA. Perhaps a written test for a "walkabout" license and a regulatory regime allowing non-criminal citations would be effective to a 60-70% mark, and more workable, though I find the idea repugnant.
As regards the big-money negligence torts, I'm happy to bite that bullet. (I'm assuming that compensation remains available as a remedy if a regulation is violated causing damage). I don't know if this is true as a matter of idea theory, but smarter regulation is a much easier ask politically than smarter tort law.
In particular, I think regulation works better than tort in practice in the contexts of motoring, pollution, consumer product safety, and medical malpractice.
There are areas like workers' comp (i.e. not common-law tort) where buck-stops-here strict liability works better than either common-law negligence or regulation would, because it is sufficiently clear who the employer is, and they usually have both the knowledge and the power to fix unsafe conditions, as well as a stake in their not being fixed if fixing them would be unreasonably expensive. Droplet-spread infections are not such an area - you don't know who the infector is without an Orwellian level of state surveillance and you don't know whether the infector or the infectee is the least-cost avoider.
If this is a proposed solution taking the real-world incompetence of governments into account, it is going to be bad. Almost everyone subject to both thinks that Anglosphere regulatory regimes work better and are less oppressive than Anglosphere tort regimes - particularly ones involving punitive damages as you suggest. This proposal replaces a complex regulatory system that enjoys widespread public co-operation with a complex tort system that will enjoy zero public co-operation (because if COVID liability policies are anything like other mass-market liability policies such as auto insurance it will be an insurance requirement that insured defendants are maximally unco-operative with plaintiffs). And the tort system needs not only to work, but also to be seen to work (because insurance companies will make decisions based on the risk of getting sued, not on the risk of spreading COVID).
Of course, you also need a system of capital adequacy regulation for vouchers which works where there is a possible trillion-dollar liability if one of your insureds becomes the index case for a second wave of the pandemic in a region that had previously eliminated it. Solvency II on steroids, anyone? Otherwise only judgement-proof vouchers will be able to offer competitive premiums.
The issue is whether people can successfully sue, and get paid. If businesses were vulnerable to getting sued if they didn't require a pass, then of course they'd be tempted to require a pass.
It's not obvious that people can account for low-probability-but-large costs when they pick jobs. You could imagine people continuning to enter such fields and 10% of them going bankrupt every year due to liability, but most entrants not noticing.Of course, that's an empirical question. My theory would predict that lots of truckers go bankrupt due to liability. If we don't observe that then my concern evaporates in a puff of smoke.
I do not see RH's supply and demand argument here. But I won't distract him by making him explain. I'll go buy a book.
Learn about supply and demand.
You realize that employees have no control over the employment contract, right. We sign what they tell us to sign or we starve. Say I’ve been appliying to jobs for months and run out of savings, and the employer sends me some contract that says “if you infect someone in the line of duty it’s on you”. Say I also happen to be a superforecaster and I realize the contract is negative expected utility. I still have to sign it, so I can keep a home. And if I’m sued I just have to put all my assets in bitcoin. Those are your options when your poor. Contracts and covid liability don’t affect our incentives in practice.
Fair enough about the distinguishing. But I think one problem when talking about such proposals is that it isn't clear just how much you are packing into the assumption that people will adopt the proposal.
I think the default assumption is to imagine that one gets to stipulate how the government behaves but then any uptake by the population at large or by corporations etc.. isn't taken for granted. This was the assumption I was approaching it from. This corresponds to a kind of assumption that the discourse you are engaged in is in some sense about illuminating potential choices for policy makers and attempting to shift what they see as reasonable.
However, I'll grant that's not the only assumption you can approach the question from but I do think it's what many people assume you to be engaged in absent other indications.
Incidentally, I have started to see companies develop and publish procedures for liability management as they plan on reopening in coming weeks, including procedures for daily attestation from each employee before arriving at work that a) they are not expressing symptoms of the virus and b) have not had contact with anyone expressing symptoms or currently diagnosed in last 14 days.
Always distinguish between if people will adopt a proposal from whether it will work if adopted. Proposers are usually making claims about that second topic, not the first.
Issues of how to make legal decisions are usually decided in the context of concrete decisions. So we can't wait until they work it out until we allow cases; we have to start the cases so they will work it out.
If this insurance is required to get a passport, then customers would be eager to buy, and so insurance firms would quickly offer products. Within a month I'd say.
If people can't afford infection insurance, that's a pretty good clue that in fact we shouldn't let them out of lockdown.
If people are doing good things but getting insurance is a cost of doing those good things, well then of course we should be willing to pay those costs, to get those good things.
The main gain we will get form this pandemic is getting better prepared for a worst one. So we should go down this path, even if we aren't sure just when we can get gains from it.
You don't really believe this would work in the current situation in the real world do you? I mean you've explained in other posts all sorts of factors that incline people to dislike this kind of system (we can't even get sufficent support to get working, liquid/uncapped prediction markets running during normal times).
With this proposal there would be huge difficulties.1) First, lawmakers would have to create rules (likely more complex than current regulations) for assigning liability (do you pay for downstream infections? This gets really complex. For instance, what fraction of downstream damage awards are you responsible for?
As a matter of pure harm any damages assessed against a downstream infection are a harm you imposed on them but that system wouldn't work for insurers (bankcruptcy lotto if you happen to insure patient 0).
What kinds of evidentiary standards will be used?
1') Do you asses damages as they do for wrongful death (in terms of lost future wages), purely as a funciton of lost QALYs plus direct economic harms)? Any system that awards higher damages to higher earners likely creates a big backlash.
2) The courts, currently unable to adequately handle their normal proceedings, would need to rapidly expand to handle the vast flood of lawsuits. (Waiting will not on spoil evidence but impose massive legal interpratational uncertainty on insurers).
3) Insurers would need to somehow read this whole new body of federal tort law, model their exposure to various legal interpratational risks (passing costs on to consumers), generate acturial models with very little relevant data, explain and market such insurance to individuals (creating the systems to store, analyze and safeguard the data needed to adequately measure risk).
3') The insurance fraud problem becomes is likely to be pretty bad as people who get sick from uncovered or unidentifiable sources see themselves as being unfairly screwed over. Unlike arson or a car crash it's unlikely there will be any evidence besides testimony of those involved. A sizeable fraction of people infected from uncovered/unidentified sources likely find covered colluder to commit insurance fraud with (they will morally justify it to themselves by saying they don't *know* that it wasn't from them even if it's unlikely).
4) We don't actually want to disincentivize people from treating the sick, delivering food or working in food kitchens. You might argue that in long term the huge insurance premiums a doctor would have to pay if they treated COVID patients but that would be an insanely large shock that (in the face of wage and price stickiness) would produce huge economic harms.
Ohh wait, that means you'd need to completely overhaul the medicaid and insurance price structures to cover massie premium that doctors would have to pay for insurance.
4') Lots of people simply won't have the money for insurance and since we don't have an extensive and reliable system that lets people take on non-dischargeable debt those people end up staying home even if they could have been net economic contributors.
—
So to sum up you need a massive overhaul of tort law, health insurance and gov health programs, spin up of an unprecedented type of insurance market, likely huge negative externalities in the resulting economic disruption as members of valuable professions like doctors get hit with huge premiums and wages/payments lag behind, likely a massive expenditure of money in both the court system and in building the new insurance system. And I'm sure I'm forgetting some.
And what do you get for all this? Insurers don't yet have the know how or ability to collect the data that would allow them to price products in a way that substantially improves efficiency (giving them that data creates another massive externality from data privacy risks). Far from a super rational system you get a bunch of fudges and guesses about good ways to adjudicate COVID liability in court.
It's a neat thought experiment but it's not even close to practically workable in this pandemic.
Your article states, "the more types of voucher-client contract terms we are willing to enforce, the more levers vouchers gain to reduce risks." For the model you propose to be effective in the context of spreading a respiratory virus vs a car accident, I'm theorizing that - in comparison to auto insurance - a) the terms must require more overhead (tangible and intangible costs) to adequately defend your level of liability; b) the variables affecting risk are not distributed evenly across classes.
Regarding that overhead, insuring against car accidents is a pretty passive policy to carry (which incidentally, arguably does have problems in pricing across classes, if not qualifying). Your suggestions for terms which could effectively insure against spreading contagion include logging movement, maintaining and logging surplus PPE supplies, logging contact points, etc. This requires ongoing activity and spend, the lack of which substantially undermine one's protection when the critical challenge comes. Those able to afford the tools and services for automating these tasks are very advantaged. Maybe the insurer could provide the means to do this though, much like Progressive Snapshot. As an unrelated aside, all of this info could be subpoenaed, so let's at least acknowledge that there's some attention required there at a later date in re adoption and compliance.
Your Pizza Risk Estimator implicates some professions over others for increased risk. Those professions with higher risks are often those at the lower income brackets, while knowledge workers able to work in isolation are advantaged. This has cost or qualification implications across classes, no?
Then the basic context here is that we're implicating not the activity of speedy transport on public roads, but the daily, required activities of basic living - socializing, walking about, buying groceries, assembling. So there's a huge incentive to dumb the barriers down as much as possible if you want compliant participation, undermining the whole point. It works for automobiles but maybe not best solution for protecting against disease.
I understand you're arguing this as an alternative to costly lockdowns, and anything probably makes more sense as an alternative to perpetual lockdowns. But there is a whole legal issue with this idea that perpetual lockdown is even a threat. The right to assemble at some point must take priority over emergency powers, whose granting must by definition be exceptional (right??).
You may counter that one of these regime models may be the required condition to lift the emergency powers. Ignoring all legal considerations, I think just from a practical standpoint (given the above reasons) a license and regulatory regime would be preferable.
In reality, you have the stronger case for an insurance and liability model considering SCOTUS has precedent authorizing the compulsory purchase of insurance as a condition for existing in the US, where they have no such thing for licensing. ;)
Auto liability insurance is not watered down to let anyone qualify, and that's the process I propose to use here. I don't see the problem of applying this "across" classes.
Fun thought exercise but you know in the US this would unquestionably be disproportionately applied across social/racial classes (reasonably or not). Imagine if you "badged" the inverse - those who are insufficient. Same system, but not very good optics. To head off your "all-or-nothing" challenge, I would note that these systems of certification and attestation work to a degree, but you'll note that the driver's license test is watered down sufficiently to allow any idiot to pass by design. How would your proposal satisfy the need for inclusivity and maintain efficacy? I roll my eyes at the very question, but in world we have , a proposal not considering this angle can be considered DOA. Perhaps a written test for a "walkabout" license and a regulatory regime allowing non-criminal citations would be effective to a 60-70% mark, and more workable, though I find the idea repugnant.
Big money torts are mostly against big firms; I'm proposing torts against ordinary people, where I presume you agree tort now works better
As regards the big-money negligence torts, I'm happy to bite that bullet. (I'm assuming that compensation remains available as a remedy if a regulation is violated causing damage). I don't know if this is true as a matter of idea theory, but smarter regulation is a much easier ask politically than smarter tort law.
In particular, I think regulation works better than tort in practice in the contexts of motoring, pollution, consumer product safety, and medical malpractice.
There are areas like workers' comp (i.e. not common-law tort) where buck-stops-here strict liability works better than either common-law negligence or regulation would, because it is sufficiently clear who the employer is, and they usually have both the knowledge and the power to fix unsafe conditions, as well as a stake in their not being fixed if fixing them would be unreasonably expensive. Droplet-spread infections are not such an area - you don't know who the infector is without an Orwellian level of state surveillance and you don't know whether the infector or the infectee is the least-cost avoider.
So you recommend eliminating all tort law and replacing it all with regulation?
Much virus reproduction is via groups of friends and family getting together for meals and similar.
If this is a proposed solution taking the real-world incompetence of governments into account, it is going to be bad. Almost everyone subject to both thinks that Anglosphere regulatory regimes work better and are less oppressive than Anglosphere tort regimes - particularly ones involving punitive damages as you suggest. This proposal replaces a complex regulatory system that enjoys widespread public co-operation with a complex tort system that will enjoy zero public co-operation (because if COVID liability policies are anything like other mass-market liability policies such as auto insurance it will be an insurance requirement that insured defendants are maximally unco-operative with plaintiffs). And the tort system needs not only to work, but also to be seen to work (because insurance companies will make decisions based on the risk of getting sued, not on the risk of spreading COVID).
Of course, you also need a system of capital adequacy regulation for vouchers which works where there is a possible trillion-dollar liability if one of your insureds becomes the index case for a second wave of the pandemic in a region that had previously eliminated it. Solvency II on steroids, anyone? Otherwise only judgement-proof vouchers will be able to offer competitive premiums.
The issue is whether people can successfully sue, and get paid. If businesses were vulnerable to getting sued if they didn't require a pass, then of course they'd be tempted to require a pass.