Firstly, the entire port of insurance is to spread risk. If the the market for insurance is "too efficient" at determining who is high-risk and who is not, then it is no longer fulfilling its social function.
Secondly, you seem to be implying that only aggressive drivers are getting flagged by the data. But the computer doesn't know if you accelerate and brake hard because you're an aggressive driver, or if your circumstances require it. Maybe you have to commute to your job in heavy stop-and-go traffic with difficult merges.
Finally, this kind of surveillance changes behavior in ways that may make things less safe. If you're driving down a street and a ball bounces out from behind a parked car do you slam on your brakes out of fear that there is a child chasing after it, or do you think that hitting the brakes might make your premiums go up so you just hope that there isn't a child coming.
> But the computer doesn't know if you accelerate and brake hard because you're an aggressive driver, or if your circumstances require it. Maybe you have to commute to your job in heavy stop-and-go traffic with difficult merges.
I would contend that this doesn't matter in terms of risk. It doesn't matter if the risk is caused by the person being a bad driver or if the risk is caused by the commute route that you take at a certain time of day.
In either case, there's risk that is shown by the data. If you are driving a route that is stressful and risky, it doesn't matter if you're a good driver or a bad driver - you're doing something that is risky.
In days of old this was done by looking at the commute distance and likely route (the insurance company has home address and work address).
> In either case, there's risk that is shown by the data
Somebody this there is risk in the data. Big difference.
In any case, you can buy insurance that does this, they install telematics box and sends them data -it’s available, and not popular. So the market has spoken
> the computer doesn't know if you accelerate and brake hard because you're an aggressive driver, or if your circumstances require it
Preface: I'm not strongly for these hyper monitoring systems of driving patterns for insurance.
Often being in a situation where you have to slam on your brakes means even if you're a good driver you're in a lot more riskier environments than a driver who doesn't often have to. A good driver that is rarely in a situation where he has to slam on his brakes probably gets into fewer collisions than the same driver in an environment where he needs to do it often. Often having to brake hard is still potentially an indicator of higher claim likliehood.
Yeah but if you really optimize it to ad absurdum, everybody will pay according to damages they will cause, negating whole point of 'socialized' spread-across-variable-population insurance.
End users shouldn't generally want insurance to run over-optimally, that benefits just shareholders of given insurance and not population overall. That I consider myself above-average driver changes absolutely nothing in this. And those stupid enough who think similarly and think they should therefore pay less can and will be easily hit from unexpected angles and end up same or worse (age, old injuries/mental diagnoses, family history and gazillion other params which are/should be mostly illegal to optimize against, and you have little to no control of).
I'm not generally for hyper analyzed metrics gathering in this case (I don't like the privacy implications), but I'm generally for riskier drivers paying significantly more. Individuals should feel the costs of their driving more in the US IMO. I don't like heavily subsidizing people who drive so recklessly.
Maybe then they'd realize overly building car dependent cities isn't that great in the end.
Convincing risky drivers to pick an option other than driving seems great to me. I'd be happy if a huge chunk of drivers couldn't drive anymore. It would make everyone safer and save a ton of lives.
Some optimization yes, but this has been already achieved decades ago by simply using driving/accident history that insurances shared among themselves probably since they went digital. All the consistently bad drivers will easily fall into this.
What I talk about is anticipation/prediction, possibly wrong conclusions from data (since we all know data can be pretty bad or incorrectly analyzed), also no way to correct any incorrectly derived bad rating.
There are no consumers winning in this scenario, even if it may feel intuitively as such if you are a stellar safe driver. Also there are many second-order effects, ie poor risky people pushed out of insurances, still driving since in US you can't do anything without a car in rural places, still causing accidents but no way in hell to pay back, ever. So we move the losses from private corporations to random citizens caught in some bad luck.
I'd say keep the risk at those corporations, they anyhow still manage to earn billions annually, no need to make their life even easier.
> if you really optimize it to ad absurdum, everybody will pay according to damages they will cause, negating whole point of 'socialized' spread-across-variable-population insurance
Except spread out over time, which is still a net-benefit.
Ultra efficient premiums should include the time domain by charging premiums in winter vs. summer based on risks.
Keep getting better and a ‘perfect’ system would bump the premiums pre accident to cover the full costs of that accident immediately before your accident. Making insurance a pure dead loss for consumers which means insurance must be inefficient to be useful.
Hot take: maybe that is actually desirable? If insurance has a plethora of good actuarial information and risk modeling, and one's insurance costs are high, that's the market telling that person that they probably shouldn't be driving. If high insurance costs promote alternate means of transportation for highly risky drivers, that's a net win for everyone on the road.
the point of insurance is to make sure if you get into a crash with somebody else and they are at fault, that they even have the means to pay you in the first place. if forced to, people make the decision to drive uninsured because it is nigh impossible to travel around this country otherwise, and then it's worse when they do get into an accident and cannot pay the other driver's costs. you can't squeeze blood from a stone.
this is also partially the reasoning behind ACA requiring health insurance; hospitals were struggling because they could not collect from people with no money that were winding up in the ER.
Sounds like we need better enforcement of insurance requirements. Driving a car uninsured should have some extremely stiff penalties. Once again, maybe if the penalties are high enough and the costs expensive enough for more people, there'd actually be more of a push to re-think overly car dependent life.
It will optimize society overall to have risky, uninsurable drivers not driving.
It's not like healthcare. There will always be healthcare costs. We don't have to have everyone drive all the time, we choose to.
> Driving a car uninsured should have some extremely stiff penalties.
I want to agree with you, but wonder if you have ever been poor? When you need the car to get to work so you can feed your kids, but you can't afford all of
- feed kids
- rent
- insurance
because you got hit by a surprise medical bill (kid got sick, maybe?)
I'm strongly in favor of your end goal (less car-dependent life), I'm just cautious about using punishment as a way to get there.
Rather than subsidize the poor person being poor via insurance, we can subsidize them by say, building public housing nearby their place of work, or direct payments.
Let's expand that. Let's name your hypothetical person, Bob. He's living on the edge financially, and decides to go without insurance.
Let's bring in another person, Alice. Alice is also not in great financial shape. But Alice is able to pay for insurance and follows all the rules. Alice has a small amount of savings, go Alice!
One day, Bob hits Alice. It causes medical issues for Alice. Alice might have insurance, but it's potentially still expensive for Alice. Because of her injuries she can't work for a few weeks. She works hourly, so now loses wages. Luckily with FMLA she won't necessarily lose her job, but she needed every paycheck. But it doesn't really matter, because her car is now gone. She can't drive to work anymore. She can't drive to groceries. She can't afford a car, as a huge chunk her savings went to cover those medical bills and missed paychecks. She's pretty SOL huh.
Sounds like we need to let Bob off the hook for inflicting all this on Alice. After all, he needed to drive without insurance.
No. We should just make it possible so Bob didn't need to drive in the first place instead of excusing his choice to still drive when he couldn't really afford it. We should structure the incentives so Bob doesn't want to drive if he can't afford it.
People driving without insurance ruin lives like Alice's all the time.
That scenario is completely missing the point. Bob has to drive because of the lack of good public transportation and the fact that he can't afford a house near his work. Not to mention if he can't afford a car with insurance it's because his job simply doesn't pay enough.
It's much more likely that external factors put him in that situation, rather than himself. Yet you propose we should punish him personally and paint only Alice as a victim. That's naive. Both are victims.
These are systemic problems and trying to solve them with individual punishments is only going to hurt individuals while not fixing the underlying issues that really matter.
I'm fully aware Bob is the victim. That's why I'm saying the solution is to make it so Bob didn't get in a car. He is a victim of car dependency. Subsidizing insurance to make sure Bob could always afford it isn't solving that. Ensuring he can always afford insurance just furthers his victimhood.
But the only question that really matters in the end is: is that profitable?
If we really cared about safety many people would not be allowed to drive in the first place. Tests would be much more strict and rightly so. But it's way more profitable to let those people spend money on cars (and eventually kill people) than it is to provide good public transportation.
Especially with the auto industry, they have basically won the lobbying game. Most people can't even imagine a world where cars aren't in the center of it, so we keep moving the goalpost...
Realistically, the other person ends up in jail and then you still can’t get compensated for car repairs. It’s not super clear why that would be better than the pooling of risk that happens today that results in compensation.
American society has chosen to create a built environment that is inhospitable to anything else except driving, so I don’t think this is actually a real choice. The people who are getting into crashes all the time are not the ones pushing for that built environment.
I'm saying they should face such extreme penalties it's not worth it anymore and have it policed so those people aren't on the roads in the first place to get in a collision. Right now you're pretty much not going to have anything happen driving without insurance unless you get into an accident that you can't escape from. If it was highly likely for someone to face judgement every time they went out you'd probably have significantly fewer uninsured drivers out there.
Then you basically just have a roving class of jobless people stuck at home.
We’re not talking about the currently uninsured. We’re talking about a “hot-take” proposal to significantly increase the amount of uninsured people today, people who today can get insurance.
Or we get a stronger political will to no longer force car dependency on everyone when people actually start to understand the real costs of car-centric design.
When the stores and restaurants where rich people shop can't get anyone to work there anymore without paying people the full costs of actually driving far away maybe there will be a will to change things.
Lots of people around the world can go to work without needing a car.
Using our justice system to meter out extreme punishment would mostly impact poor people, and it is not as if the US has historically been very good at listening to the suffering of poor people.
The Bay Area can barely keep service workers in housing and its taken several decades of this problem to get even the slightest bit of progress.
Being frank, you shouldn't be making the 40+mile commute in a car if you can't afford all the ramifications of it. It's a failure of societal design to force people to drive 40 miles on a vehicle they can't responsibly afford to operate just to survive.
In the end, they shouldn't live 40mi from where they work. It's not a good thing to force such a lifestyle.
Honestly it's depressing you're suggesting we should continue to force people to spend so much of their productive lives commuting to dead-end jobs that will never lift them out of the poverty of their situation. It's sad you're continuing to argue people should live an hour+ away from where they work, and that should just be the norm and the basis for our designs.
If you want to live 40+mi from your work and can afford all that involves and are willing to live with the tradeoffs, sure go ahead. Pay the tolls for the highways. Pay the congestion fees. Choose to spend more time with the insides of your car than you do spending time with your family on an average weekday. Pay the higher insurance compared to those who live close or take the train. Just quit asking for handouts and subsidies to pave over other people's homes, force bullshit parking minimums which lead to seas of empty pavement, demand other people pay for the roads you drive, etc.
your proposed change will take decades to rework cities to move away from car-centric city design, to introduce public transportation, to rework current districts, to move shopping malls/restaurants closer to living districts etc.
all of this just because you wanted to make stricter insurance just to make it more (by how much?) efficient for insurers, so that they would make more profit.
No, I'm suggesting it because I don't want to pay for Bob to smash several cars a year. I'm saying it because there's lots of people who have no business being out on the road. Having Bob drive when he's a bad driver makes everyone around less safe.
But I guess you'd have society pay for all the cars Bob ends up destroying. We'll subsidize him crashing cars over and over and hurting Alice but we just can't seem to find the money to add another bus line!
> If the the market for insurance is "too efficient" at determining who is high-risk and who is not, then it is no longer fulfilling its social function.
Then we need to break whatever that social function is away from the umbrella of "insurance." Mandatory car insurance is predicated on the fact that you can pay enough on average to cover your damages to others but might not be able to do so in the worst case. If we're in a world where somebody's driving exceeds the external damage bounds they can afford _even on average_, subsidizing those people is no longer the job of insurance.
Mandatory car insurance is essentially a welfare benefit. The benefit is just administered by insurance companies, because they are believed to be more efficient than the government. (This is pretty common for all kinds of benefits around the world.) Mandatory insurance is not a private contract two parties have reached at their own initiative, and the usual business considerations don't apply. If an insurance company wants to provide mandatory car insurance, they must follow government policies, not their own policies.
If someone drives recklessly and causes excessive damage, the government has other tools beyond insurance premiums. They can, for example, revoke the license and confiscate the car. Or issue a fine or put the driver in prison.
Mandatory car insurance, in theory, moves money from drivers (proportional to the risk they represent) to victims (proportional to the damages caused), plus a cut for the insurance company.
I'm missing something in your flow of ideas. Is it a welfare benefit because that idealized theory of insurance is itself a welfare benefit or because when insurance is mandated it doesn't live up to expectations?
The insurance is mandated, and the focus is more on guaranteed benefits to the victims than on fairness for the people paying them.
If you can't afford to pay for the damages you caused, that's your problem. You took the risk, and now you face the consequences. There is no reason to make the insurance mandatory to prevent that. But if the victim doesn't get any compensation because you can't afford to pay, that's a public problem. Now there is a reason to make the insurance mandatory, and it's particularly important that the riskiest drivers have insurance. If you make it too expensive for them, they may choose to drive without insurance. Which is exactly what you wanted to avoid in the first place.
> If we're in a world where somebody's driving exceeds the external damage bounds they can afford _even on average_,
Does this have any meaning at all? I think you are too busy thinking about cars and not people.
If a pedestrian is hit and crippled for life, the damages are more than you can afford for majority of the population.
Many accidents are random chance, caused by factors that cannot be controlled (weather, random technical failure, etc) and lethal accidents do not tend to repeat.
I think we agree about the nature of accidents -- they have some chance of occurring, the damages are often more than any one person can afford, and beyond a probabilistic assessment you have no way of knowing exactly how many wrecks will happen or what the damages will be.
My point is that you _can_ analyze those accidents probabilistically. Somebody who practices defensive driving, never drives over 25mph in a residential, only drives in broad daylight, and only travels 500 miles per year will have a very different baseline, both in number and severity of accidents, than somebody who habitually blows through residential stop signs at 60+mph and drives 50,000 miles per year.
Insurance concerns itself with flattening those spiky probability distributions. The first person will likely never severely injure a person even if they get in an accident, just from the difference in miles driven they're going to be in 100x fewer accidents per year, and probably much better than that because of their other safety practices (call it an additional 2x factor).
Just to have some hypothetical numbers to play with, the safer driver has an average of 0.001 wrecks per year, and the average damages might be $20k (fender benders, minor hospitalization, ...), so they have about $20/yr worth of risk to insure against (yes, I know you have to integrate over probability of different types of wrecks or whatever; this is a simple HN comment with ballpark numbers). The dangerous driver has 0.2 wrecks per year, and the average damages might be $100k (total both vehicles, major hospitalization, ...), so they have $20,000/yr to insure against.
_Insurance_ concerns itself with factoring in those relative risk profiles (along with the time value of money and whatnot; it gets a bit complicated) and guaranteeing that the individuals only have to pay their flat premiums (and optionally a flat deductible per wreck, though a decently high deductible is a good idea for most people) instead of risking bankruptcy and then some (it's the "and then some" that made car insurance mandatory in the US -- ensuring that the unlucky person you crashed into can still be made as financially whole as possible).
And that's all I meant with the "even on average" comment. If the safe driver's premiums were a bit more than $20/yr, and the dangerous driver's premiums were a bit more than $20,000/yr, that would be _fine_ from an insurance perspective.
The person I was replying to was talking about the "social function" of insurance, and my real point is that if the social function is to allow that dangerous, more expensive driver to continue to drive (not necessarily as bad of an idea as it sounds in the abstract -- if they couldn't legally drive, would they do something more dangerous or less insured instead? is it worth the additional costs we inflict on people when driving is nearly mandatory but we just won't let them?), _insurance_ isn't the tool to enable them. We should be thoughtful and explicit when providing those sorts of subsidies, instead of hiding and burying them in a tangentially related financial instrument.
Mind you, going 1-2 comments up the chain and referring to the data collection, I still think that's bad for other reasons, and using "too complicated" of models (under the assumption that they'll likely never be analyzed by a real person and have a chance of being egregiously wrong) isn't great either, but pricing insurance based on what you know about a person isn't bad in and of itself.
> Firstly, the entire port of insurance is to spread risk. If the the market for insurance is "too efficient" at determining who is high-risk and who is not, then it is no longer fulfilling its social function.
Not necessarily! If you’re a physician, you have liability insurance against malpractice claims. If the insurer can correctly determine that you are a particularly incompetent physician and your premiums should be so high that you’re effectively just pre-paying the future settlements for your inevitable future malpractice cases, you will not be able to afford those premiums and will be forced to stop practicing medicine. In this scenario, the insurance market has satisfied its social function.
In principle, incompetent and unsafe drivers could similarly be identified and priced out of being able to drive, and the act of doing so would serve a purpose of making the roads safer for everyone else.
I think these systems are bad and I think vehicle manufacturers piping data directly to insurance companies should be illegal whether you consent or not. That being said, "but privacy" has always struck me as a weird argument.
There is no reasonable expectation of privacy when a vehicle with a particular license plate is at a particular intersection or on a particular public street. Some jurisdictions extend some privacy rights to whether or not a particular person is driving it but many don't.
So when it's in public, the location of your car is not "private" information. How is acceleration, braking, or other engine telemetry data private?
The right to privacy is not boolean. There is a spectrum of reasonable expectations of privacy. Also, the right to privacy in public is not necessarily zero. The varying degree of rights to privacy in public are less than in private settings, but often greater than zero.
While your insurer may, by happenstance, see your license plate while you are driving around, they certainly are not following you 100% of the time. The degree to which your location is tracked when your license plate is plainly visible, is different than degree of privacy expectation when that location is being automatically recorded.
> So when it's in public, the location of your car is not "private" information.
It's just not that simple. It would be nice and convenient if "private" was a boolean value. But it is very much a float type:
> they certainly are not following you 100% of the time
I don’t follow the argument being made here.
The state of modern surveillance is different from what is described. People are consenting to precise monitoring for slightly lower insurance premiums. The question is: should it be illegal (should we fight against monitoring) for some moral/philosophical reason, or are we accepting it because that’s the way things are.
There is not necessarily informed consent with the type of informatics data collection described in these suits.
This isn’t a lawsuit about a “progressive snapshot” type of device.
It has been described by many that these vehicles had been opted in at the dealer with little to no notification, and the data subsequently sold to brokers, who then sold the data to insurers.
If I personally attach a GPS tracker to your car or follow you everywhere with a camera, I will get charged and thrown in prison despite it all being in public. This behaviour will be considered stalking and threatening.
People on here argue about laws as if they are universal, but if you actually try to do the exact same thing as these companies do, you will find out the difference very quickly.
Insurance spreads risk over outcomes. The social function of spreading risk over people (subsidizing risky behavior) is both incidental, accidental, and a clear moral hazard. Insurance would still be extremely useful even if we knew with perfect certainty the probability that you would have a car accident in the next 5 years.
Seems dangerous when the "risky behavior" is outside the control of the person being insured. Or if what you infer as risky behavior isn't actually risky. Is being born with a heart defect risky behavior? We have regulations to avoid discriminating over pre-existing conditions for a reason, and I could easily see these methods leading to similar bias. Doubly dangerous when insurance across multiple industries are essentially or literally required.
It'd be easy if we had a magic ball that could pinpoint risky immoral behavior, but I don't think we're anywhere near there yet. I barely trust the sensors in these cars to not report bad data, let alone trust the company to not extrapolate cynical conclusions that help their bottom line at the expense of people's wellbeing.
>Firstly, the entire port of insurance is to spread risk. If the the market for insurance is "too efficient" at determining who is high-risk and who is not, then it is no longer fulfilling its social function.
Insurance is a business, where the insured pays an insurer to pay for costs for damages that the insured cannot afford to pay. The insurer's job is to calculate the amount of premium necessary to ensure they can afford to pay for the damages, but they also have to sell the insurance at competitive prices, so they probably have to calculate premiums based on the risks that each insured represents.
>Secondly, you seem to be implying that only aggressive drivers are getting flagged by the data. But the computer doesn't know if you accelerate and brake hard because you're an aggressive driver, or if your circumstances require it. Maybe you have to commute to your job in heavy stop-and-go traffic with difficult merges.
The goal is not to label drivers as aggressive or not aggressive. The goal is to tease out the factors that lead to claims. With a sufficiently large dataset, it should be possible to tease out whether or not characteristics such as stop and go traffic is an indication of higher likelihood of claims. If an insurer does not calculate this property, then a competing insurer probably will and hence be able to offer lower premiums to less risky customers.
>Finally, this kind of surveillance changes behavior in ways that may make things less safe. If you're driving down a street and a ball bounces out from behind a parked car do you slam on your brakes out of fear that there is a child chasing after it, or do you think that hitting the brakes might make your premiums go up so you just hope that there isn't a child coming.
If a single event of slamming brakes is causing premiums to go up, I doubt that insurer is pricing premiums accurately. However, if the insurer experiences greater losses in neighborhoods where people are slamming their brakes more often, then obviously the risks are higher in that neighborhood and premiums need to reflect that.
The higher premiums themselves would tell people in the neighborhood that either they are driving too fast, or they are not looking after the children. And if people are choosing to risk plowing into a child for the sake of their premiums, then that is more of a moral quandary than a problem with insurance pricing.
If the goal is to subsidize a specific population, then that should be a government function via taxes.
> If the goal is to subsidize a specific population, then that should be a government function via taxes.
As I understand, all 50 states require insurers to write high risk drivers policies through assigned risk plans. I guess you could shuffle the money through the state as well, but it seems like an inefficient way to accomplish the same.
But really, car insurance and their respective mandates exist to protect the innocent, not drivers themselves. e.g. No state requires that you insure your property.
But yeah, that’s kind of my point. No state (except NH) wants people to be able to run others over, default on any damages, and leave the innocent person screwed. That’s why liability insurance is mandatory, and it’s why (mostly) first-party insurance is not. It is a social safety net.
> Firstly, the entire port of insurance is to spread risk.
I think that definition is incomplete: Insurance is to spread risk equitably.
That means accounting for disparate probabilities and disparate impacts. For example, consider "house burns down" insurance, where premiums depend on whether the house is/isn't near a wooded area, and whether it is a cheap/expensive house... And yes, also whether or not the homeowner has a passion for homemade fireworks.
> If the the market for insurance is "too efficient" at determining who is high-risk and who is not, then it is no longer fulfilling its social function.
While I agree that various dystopic outcomes are possible, the problem is not better knowledge about risks itself. Improved information about the dangers we're trying to avoid or fix is--all else being equal--always a good thing.
The real problems stem from those other no-so-equal factors like:
1. Imbalanced power relationships. (Strongly implicated in the rest of the list.)
2. Opaque decision-making that cannot be reviewed or appealed.
3. Information not being fairly discovered/shared. (Customer hides known higher risk, insurer hides lower-than-expected risks to squeeze out more profit, etc.)
4. Bad contracts which pull the rug out from under people because of how they handle changes in knowledge even when risks haven't actually changed. Imagine health-insurance which covers Giant Monsterification, but later a test reveals patient has Godzilla genes, and now the customer is dropped... Even though the originally-covered probability itself hasn't changed, only our knowledge about it.
> Insurance is to spread risk equitably. That means accounting for disparate probabilities
Everyone defines equitably according to their whims. Does it mean the riskier pay more? Those can afford it pay more? The unsympathetic pay more? Et cetera
While it's true that Solutions Are Hard, that doesn't change the fact insurance is absolutely not meant to spread risk blindly.
Even early proto-insurance of seagoing merchants mutually carrying one-another's trade-goods on their various ships (in case one sank) still cared about how some cargoes were more valuable and some ships were more seaworthy.
>Firstly, the entire port of insurance is to spread risk. If the the market for insurance is "too efficient" at determining who is high-risk and who is not, then it is no longer fulfilling its social function.
This might be definitionally correct, but that doesn't mean it's right.
There's no reason you shouldn't rewarded for being a better-than-average driver (or conversely, punished for being worse).
>Maybe you have to commute to your job in heavy stop-and-go traffic with difficult merges.
You don't think we have the technical ability to determine this, to some degree? We have self driving cars for crying out loud.
>If you're driving down a street and a ball bounces out from behind a parked car do you slam on your brakes out of fear that there is a child chasing after it, or do you think that hitting the brakes might make your premiums go up so you just hope that there isn't a child coming.
What a reach.
You hit the brakes, because if there is a child, and you hit them, your premiums will be even higher.
> You don't think we have the technical ability to determine this, to some degree? We have self driving cars for crying out loud.
I actually work on self-driving cars, so I have some experience on this. Trying to predict safety performance based on more easily measured metrics like hard stops is hard. AV companies spend a lot of time thinking about it. I don't think they get it perfect either.
Secondly, you seem to be implying that only aggressive drivers are getting flagged by the data. But the computer doesn't know if you accelerate and brake hard because you're an aggressive driver, or if your circumstances require it. Maybe you have to commute to your job in heavy stop-and-go traffic with difficult merges.
Finally, this kind of surveillance changes behavior in ways that may make things less safe. If you're driving down a street and a ball bounces out from behind a parked car do you slam on your brakes out of fear that there is a child chasing after it, or do you think that hitting the brakes might make your premiums go up so you just hope that there isn't a child coming.