Hacker News Reader

Comment by taftster | Hacker News Reader

And here's the thing. These types of driving habit trackers (not necessarily new in the auto-insurance industry) are touted as a means to reduce the cost of insurance for good drivers. But what qualifies as a "good driver" is likely to be more like 10% of drivers. For everyone else, this won't reduce the insurance costs at all.

It's frankly just a way for insurance companies to extort more profit from drivers. I don't want a black box in my car of any type, even if I already qualify for the best possible insurance rates. Your driving habits are a huge privacy concern and we should all guard them very carefully.

View on HN

I wonder how insurance mandates interact with this trend. When the state requires insurance, but insurance requires disclosure of all movements, what recourse is there? It seems the only party with an interest in privacy would be the purchaser who just wants to drive to work.

This sort of "outsourcing of standards" is really disconcerting.

The state is likely to see this as a positive anyway. Insurance companies can bill this as a major boon to law enforcement, traffic regulation, parking violations, and of course think of the children.

There will always likely be an option without the black box. It will get more expensive though, as drivers self-select into different groups.

> When the state requires insurance, but insurance requires disclosure of all movements

Not necessarily. We do some driver behaviour analisys and it accounts for harsh acceleration/deceleration on some speed intervals and harsh cornering. No other tracking is nedded. One could standardize a set of metrics and give the driver a mark. But it's still discrimination by algorithm without any kind of appeal in the end. If you analyse the data from the GPS for instance, any GPS reflection (eg. tall buldings) could cause a jump in speed and trigger false alarms. Same with the accelerometer but on a lesser extent.

Point well taken, but it’s still another case of “we’ll only use these data, not those, trust us.” Even if the insurance corps are truly trustworthy, it only takes one mistake to have stored data leak. After the last experience I had with USAA after a drunk hit my parked car, I would personally very much like to opt out of trusting in this way. I hope that remains an option.

Try making a left turn in a state with sub-par signals without tripping the accelerometer threshold for a bad driver. This will punish good drivers forced to deal with bad civil engineering.

Will the >insurance< be >punishing< the drivers?

Or is it more accurate to say it gives them an incentive to address the bad civil engineering through their behavior.

The insurance company is operating on an actuarial table that shows cornering with those accelerometer values is statistically involved in more accidents. The user will slow down because now they're aware of the requirements to make a safer turn. That is a third order effect based on the insurance company adding more information to the driver's calculus.

The 'punisher' here is the design that forces you to choose between dangerous turns or slow navigation.

So traffic gets worse. Got it, thanks.

By the way, I guarantee their actuarial tables don't know whether you're creating high g forces in or out of control.

And now we've created a system that punishes the driver that brakes hard to avoid the tentative turner, who is incentivized to make the situation MORE dangerous by turning timidly.

I reject the certainty with which you stated your opinion. Actuarial tables don't "know" anything, but they are models, and as such, they can include in control / out of control as features.

I think it'd be pretty easy for sensors with a CPU to determine what's out of control. That tech is well developed and very stable, similar to how iOS can tell the difference between me walking and me shaking my phone.

Sensors can't tell what is appropriate. Is a 1.5G acceleration automatically dangerous? Maybe it's necessary, given the circumstances. Circumstances that are never considered when the insurance company chooses an arbitrary number that conveniently boosts their revenue.

That's correct.

Because they're not trying to reward you/punish someone else.

It's a risk mitigation tactic, despite what Progressive's commercials frame it as.

Insurance companies already stratify premiums. You can be a perfect driver and pay extreme premiums because you happen to be young and live in a certain neighborhood rather than another, since actuaries are forced to rely on correlations rather than causation since they lack this data.

As someone who has never gotten into a wreck nor even gotten a speeding ticket, I for one would welcome a data driven approach to pricing premiums rather than having me binned with my peer group who has no qualms texting and driving drunk over the speed limit, on average.

Fair point about the bins and a "data driven approach" to premiums. But I think you overestimate how much "you could save" by giving away your driving habit privacy. Most auto-insurance trackers are currently offering discounts "of up to 15%" if you qualify as a "good driver". Seriously, that's like what $40 on a typical 6 month policy? And you have to drive perfectly at all times, no speeding ever, no rapid acceleration, no sudden stops, etc. All of that is being tracked. And you most definitely cannot get in an accident at all.

Even with your good driving habits, you are still subject to randomness equally on the road at any given point in time. Your demographics and "bins" aren't going to make the roll-of-the-dice any different for you. Driving is a random series of events and you can be caught in a situation that is your fault (legally) just the same as anyone else on the road.

This move is nothing more than a company trying to make extra money by selling your private data. It's the new business model of the information age, and all traditional companies are looking at new ways to extort your privacy for the sake of profit.

> But I think you overestimate how much "you could save" by giving away your driving habit privacy. Most auto-insurance trackers are currently offering discounts "of up to 15%" if you qualify as a "good driver". Seriously, that's like what $40 on a typical 6 month policy? And you have to drive perfectly at all times, no speeding ever, no rapid acceleration, no sudden stops, etc. All of that is being tracked. And you most definitely cannot get in an accident at all.

This all seems like conjecture. The auto insurance market is pretty competitive, and if an insurance company misprices their insurance, there are plenty of other insurance companies out there that will correct it.

If insurance companies just wanted to raise rates, they already would have.

I'm pretty sure the state insurance commissions have something to say about that, too.

>Seriously, that's like what $40 on a typical 6 month policy?

Interesting that you think car insurance costs around $500/year. Claims that people pay close to triple that on average seem to be abundant.

https://cars.usnews.com/cars-trucks/car-insurance/average-co...

OK, fair to call that out. My bad for grabbing numbers from my arse.

So let's take the average, which per your source, says $1400 / year (or $700 per 6-months). $700 * 0.15 = $105 savings per 6 months ($17.50 per month).

So now you need to ask, will you qualify for the full 15% discount? If you stop too quickly or roll through a benign stop sign, you can forget it.

And worse, don't you think that the insurance company in question already raised their rates by 15% only to give you a chance to earn that back with (quote) "good driving habits"? If they ding you for anything, then you're going to pay the higher rate.

It's like a grocery store member card. If you don't have it, they just don't care, they just charge you the non-member price. But their member price is already profitable, so the few shoppers coming through and buying without a membership are just paying an additional gravy to their profits.

Insurance companies want to get into this game, desperately. Oh, you don't have their tracker app installed? Well, you're going to pay the non-premium expensive rate and they get the keep the profit. Your good driving is not going to offset your demographic either.

And then there's all the privacy concerns. No thanks, I'm going to opt out and not buy into the $18 per month non-discount with some gimmicky insurance company and instead give my business to a company that doesn't want to play premium discount games.

>Driving is a random series of events and you can be caught in a situation that is your fault (legally) just the same as anyone else on the road.

This is true, and that's why you pay more for car insurance in Los Angeles than in the desert. There is a base rate for the given area that everyone pays premium to, given the danger of a given square foot of road is different for every area. This is already all quantified and baked in, even driving habits from indirect sources like google maps or road conditions, so personal data collection will not change this.

So why data collect at all if so much variance is due to local danger? Because a huge component of the variance is also explained by drivers skill, and insurers proxy this based on correlations from your demographic data, since they lack any other datapoint on your particular driving habits (other than accident history).

This is why powerful sports cars with a twitchy throttle that are hard for most people to control cost more to insure than a minivan, and why young people pay far more for insurance than older people, and why males pay more than women for insurance as well. These people might operate over the same square footage of road and experience the same risks, but pay sometimes vastly different premiums.

If you were an F1 driver who had split second reaction times and could control a vehicle better than anyone on the road for probably 1000 miles around, certainly you'd like for insurers to take this into consideration before they bin you with the other 23 year old male Los Angelenos who drive a ferrari downtown, probably a rather reckless group of peers if we had to stereotype.

>you overestimate how much "you could save" by giving away your driving habit privacy

It amuses me that when it comes to phones tracking us, apps tracking us, google maps tracking us, etc.. we don't get the same outrage.

Also, the answer to your concerns is regulation. Contact your state reps.

I'm actually okay with pricing based on driving habits. Sure it won't be perfect, but it seems more just than pricing based on whether you're married or not or have a good credit score like insurance carriers currently doing.

Agree. I might not like it for myself. But I benefit knowing there’s a cost for all the horrible drivers out there might make them drive more safely.

Sorry others are downvoting you, I know exactly what you mean. My SO and I are age 24 living in a relatively high-risk area for theft and accidents. But we both live within walking distance of work, and since covid hit, we're not driving hardly at all. Our insurance doesn't seem to care.

We pay close to $300/mo for insurance. What a crazy prerequisite to participate in modern society. Imagine if all that money went towards building a train/subway network instead.

Some quick Googling tells me there are ~2.5M registered vehicles in America, and the average cost of car insurance is ~$1,400 / year. That's $2Bn/yr we could be spending on public transit, even ignoring the costs of gas and car purchases.

EDIT: Found an article [1] claiming that personal transportation costs accounted for $1 TRILLION in spending in 2018.

[1] https://www.itdp.org/2019/05/23/high-cost-transportation-uni...

You have the wrong insurance maybe? Geico recently refunded money to its customers because of the change in driving habits during Covid quarantine. That seems to indicate a company that at least balances their insurance needs based directly on how many miles are being driven.

We have Progressive because they gave us the best quote when we signed up (neither of us had a car insurance history.)

We look every couple months or so for a better quote, but so far, nothing. Geico would have us paying even more.

Geico refunded customers because they were required to by law as many areas limit the maximum profits insurance companies can make.

> Some quick Googling tells me there are ~2.5M registered vehicles in America

There's no possible way this is correct. If you assume 1 car per 2 people, you get 164M registered vehicles. Google suggests the actual number is 285M.

Though many insurance policies cover multiple vehicles, and that's probably $1,400 / year / policy not $1,400 / year / vehicle on insurance.

> For everyone else, this won't reduce the insurance costs at all.

It will likely raise it to subsidize things, somebody else has to pay your premium... Companies are never in the business of making less money.

"Companies are never in the business of making less money"

That's just a perfect way to word that. Gonna save that one for later.

>It will likely raise it to subsidize things, somebody else has to pay your premium

I don't know what this means, but an insurance company can be incentivized to lower premiums as a means to win new business. If an insurance company offered me lower premiums in exchange for monitoring my driving, I might do it. I already have dash cams to ensure I don't get found at fault if it isn't my fault.

Insurance companies have been offering just this deal for years - they will stick a GPS logger into the OBD2 port on your car, and then ask you to let them pull data off it periodically[1].

1: https://www.forbes.com/sites/adamtanner/2013/08/14/data-moni...

Root Insurance does this via mobile app. They might be available in your area. Disclosure: I work at Root

In my state, for example, insurance companies required to pay out a percentage of premiums. With the reduction in miles driven across the risk pool I've been receiving refund checks.

It will raise rates for those who are already risky drivers, but will not save money for typically safe drivers. And worse, this will push risky drivers to even more abandon auto-insurance entirely. The good drivers will be forced to carry more uninsured motorist coverage, simply because the for-profit insurance companies decided to drive their bottom customers out of the insurance market entirely.

"Somebody else has to pay your premium" is really just going to manifest itself as an increase in uninsured motorist premiums. Sadly.

And once they get enough people to sign up, they can punish everyone else by putting them in the "bad driver" bucket.

It seems to me like the brackets should actually be the other way around:

90% of "good drivers" "save money" (all relative, of course), while "bad drivers" who cost a lot will pay way more.

Your smartphone already knows your driving habits. Mine likes to remind me where I parked my car. (It knows because I connect it to the Bluetooth music player).

I honestly think regulation is the only solution here.

Smartphones might know your driving habits, but is big tech selling your information to insurance companies? That would be a huge breach of privacy and most definitely would result in massive class action litigation against the tech companies.

Your driving information is one of the most private things you can protect. Did you just make a stop by the dispensary? Or park at the cancer oncologist? Or drive into a protest area in your downtown city?

I agree that regulation is needed quickly in this area, but at least in the United States, it's unlikely to happen anytime soon. We need to at least be aware and not so complacent to the changes that big business are trying to do to make additional money from your private life.

Insurance isn’t without competition. If a group of customers are being overcharged relative to their measurable risk, then that’s an opportunity for a competitor.

Correct. It's enhanced risk pooling. It's not extortion of one company uses it, it's absolutely extortion when an industry requires it, and states regulate it be adopted.

Like all opportunities to make money it will become a scenario of ‘good drivers’ get to keep the same cost (no discount) and ‘bad drivers’ (the other 90%) get to pay more.

Yes, exactly. There's no winners here except the corporations and shareholders. And it's all at the expense of personal liberty and freedoms afforded by driving.

Insurance trackers can also increase your premiums. If you drive like an asshole, the agreement states they will raise your rates.

I'd argue that most people drive like an asshole at least some of the time, which is exactly the problem. From an insurer's point of view, they want to find ways to qualify you as an asshole and therefore extract more money from you.

These trackers are literally looking at things like aggressive acceleration, sudden stops, lane changes, slightly over speed limits. All of these examples can very much be "safe" driving habits, required to safely move your vehicle in traffic. But any single event, without the context of the traffic surrounding it, can brand you as that asshole. These apps don't care about the context of the traffic around you.

>From an insurer's point of view, they want to find ways to qualify you as an asshole and therefore extract more money from you.

If an insurer starts needlessly penalizing people for minor violations of certain parameters, then another insurer will step up and offer lower premiums. I'm sure there's plenty of smart people at insurance companies that know a single event doesn't make a person more likely to cause a loss.

Not just driving habits, but timing and locations visited, hence people met, by interpolation. The black box is by all means a bug. Also being all of them closed source, there's no guarantee they don't contain circuitry to listen to conversations or hidden cameras. Paranoia? Yes, totally. Today's world fully justify that.

>I don't want a black box in my car of any type, even if I already qualify for the best possible insurance rates

This seems like an odd statement to me. Of course you don't want a monitor, if you qualify for the best rates. The people who will go along with it are those currently paying the highest rates.

Fair of you to pick up on the oddity of that statement. I didn't put that out there very well.

Agreed, a person with good rates doesn't want this. But that doesn't make a person with bad rates desire it either? Does this really improve their situation any?

Maybe there's a very small percentage of drivers that will try to prove they aren't a bad driver. This (young/male/minority) driver has been boxed into a high risk demographic and wants the chance to prove they deserve an individual rate different than their cohort.

However, what will really happen is that rates will just stay the same for them (they are still afterall in the same demographic), but additional revenue will be extracted from those that might be more aggressive or take more liberties when they drive.

I don't believe you're saving money for anyone with this scheme. All you're doing is adding additional tax to those drivers that might drive more risky. Why should the insurance companies profit any more? They need their stock prices to keep rising, at all costs, particularly at the expense of the clients needing the most insurance product. This is just another scheme to keep profits rolling up, no company wants to charge less to anyone.

The demographic boxes that we are put into for premium rates can be said to be unfair. I don't argue that point. My point is that there's nothing here that actually saves anyone money.

But overall, the more concerning part of this is the loss of privacy and freedom associated with the collection of personal driving habits. Did you drive through a part of town that is notoriously high in crime? Did you make a stop at a liquor store? Did you drive to a political rally for the opposing party?

All of these are deeply concerning questions from a freedom, privacy, liberty point of view. Big brother shouldn't have this much tracking power, and neither should big business.

Agree, even if it saved me significant money on premiums, I don’t want a third party tracking my location and driving habits. Especially when the third party is probably gonna store the data insecurely and at some point bound to get breached / hacked. This is utter none-sense.

Do they even need to get breached? Or will somebody look at that service and say "You know, this detailed geolocation data we've collected about all of our customers could be a profit center" and start selling it?

They've got more detailed information of exactly where you are at exactly what time, which gives them a competitive advantage for creepy stalking over the CA DMV: https://news.ycombinator.com/item?id=24210623

I just willingly put one of these tracker in my car. I barely drive anymore due to Covid and USAA is attempting to charge me 180 per month. Metromile lets me pay 70 dollars a month + 12 cents a mile. No brainier decision at that price point unfortunately.

It might be a good deal for you. But I wouldn't take the "no brainer" argument here. Driving habits are highly personal and you most definitely should resist being profiled by your insurance company, even for the extra break on premiums. That may be a good deal for you now, but don't count on it for long, I'm afraid.

I already am profiled by my insurance company. Only their profiles are based on correlations rather than a direct line of causation since they lack any data on me beyond inference from the population. As a result, I'm shackled to suffer from the behavior of my peer group, so if I'm young or live in a city I pay more money no matter if I am one of those perfect drivers who goes a lifetime of driving without an accident or ticket.

USAA has been returning premiums to customers during Covid. And I'd pay more to not be tracked.

Just for the hell of it, I once got a quote from Metromile when I was hardly driving at all, and it was HIGHER than my USAA rate.

Rates are regulated by the states so the mileage will vary by state.

Also, I feel like this is much fairer than the current practice of rating based on credit score.

On the flip side, it's a way to safely offer more competitive rates and outdo your competitors.

Whats even worse is if Amazon gets all your driving data to toy around with and advertise to you based on places you visit and other things it may know about you by your car habits + your online habits. Some sites do feature Amazon ads.

To play devils advocate, you don’t need insurance if you’re driving on private roads. If you’re driving a vehicle on public roads what is the expectation of privacy? Is a pilot’s privacy being encroached when they have to file a flight plan or use a transponder?

There's a huge difference between not having an expectation of privacy and having a positive expectation of stalking.

Even if you are driving on public roads, there is still a certain expectation of privacy.

Reply on news.ycombinator.com