Greetings from sunny Santa Cruz, where traffic moves at an average speed of about 5 mph on the freeways (took 2.5 hours mid-day to drive down here from Berkeley, 75 miles away). One fellow here said that car accidents account for 4 percent of U.S. GDP and that the cost of an accident was the single largest component of the per-passenger-mile cost of driving.
Thoughts? Sources?
http://www.nhtsa.dot.gov/portal/nhtsa_static_file_downloader.jsp?file=/staticfiles/DOT/NHTSA/NCSA/Content/PDF/810837.pdf tells us that approximately 43,000 Americans were killed by car crashes, but doesn’t calculate the cost.
How much is your car insurance?
To a first approximation, this should reflect the cost of accidents. Maybe there are income numbers for insurance companies, somewhere.
US GDP was $13.13 trillion in 2006. 4% of which would be $520B. That would divide into ~$1700 in accident costs for every single US citizen (300M) per year. However, the BTS estimates 6,159,000 accidents per year (http://www.bts.gov/publications/national_transportation_statistics/html/table_02_17.html), so I’m not buying the 4% number (>$2M/accident).
It seems like a plausible figure. Start off with insurance costs (no accidents=no need for insurance) assuming everyone had fully comprehensive insurance including a courtesy car, with no excess (that would be “deductible” in the US, I think), and then add all of the uninsured/uninsurable costs — traffic delays caused by accidents, deaths and injuries, time spent sorting out repairs, etc.
Incidentally, 43,000 deaths per year is a shocking number. The UK has almost precisely 1/5 of the US’s population and has about 3,000 traffic deaths per year, so less than half as many per head of population.
$1000 per year is a good ballpark figure for insurance. GDP per American is about $45,000 per year. Not every American has a car, so it is looking like less than 4 percent. On the other hand, insurance does not cover a lot of the lost work time or medical bills. If you crash your own car into a tree the medical costs will be paid by your health insurance, the Federales, or out of pocket. The time and hassles spent dealing with body shops are not figured in. The deductible of $500 or so is not figured in.
A typical economic estimate of the value of a human life in the USA is $5 million. If car accidents kill 43 thousand Americans every year, that’s a cost of $215 billion. So the lost human life alone costs around 1.6% of a $13.13 trillion GDP.
There are about 243 million registered passenger vehicles in the USA according to wikipedia. Assuming $1000 insurance per vehicle, accidents cost $243 billion per year in damages to cars. So that’s another 1.9% of GDP.
That’s 3.5% of GDP right there, before counting secondary costs such as non-car property damage, police activity, and increased congestion. So 4% sounds about right.
Mike, I’m curious what is the accident rate in UK per mile/km traveled? If I read the powerpoint correctly in the US I have to drive for about 100 million miles to die in a car crash. If I buckle up and take it easy with booze I could probably last 200 million miles or more.
43,000 deaths – still insane.
Marriagedestroyer, There are 33 million vehicles on the road in the UK, 26 million of which are private cars. Average annual mileage per car is 10,000 miles; let’s say it’s three times as much for the other vehicles and we get total vehicle miles of 470 billion. Divided by 3000, that’s a bit over 150 million miles between fatal crashes.
Statistics for 2006:
– Number of people killed in automobile accidents in USA: 43,000
– Number of people killed in terrorist attacks in USA: 0
Where then is the ‘War on Bad Drivers’ or the ‘War on Dangerous Oversized Penis Extension SUVs’?
I’m “one fellow here.”
One of the reports is the hard to find DOT HS 809 446, it puts the number at $230.6 billion, which is only about 2.3% of the year 2000 GDP. I will have to hunt a bit for the study that put it at 4% . There are 6 million accidents, and 43,000 fatalities in a year.
Of course there are some subjective numbers in the calculations, ranging from the value of a human life and the cost of traffic delays. “Accident infrastructure” (police, fire, ambulance) can be calculated to some degree but we would still want a lot of it to be around even without the accidents. The size of auto insurance payouts can tell us the cost of what they cover (collisions for yourself, liability for others) but that’s only part of the story.
The context of all this is just how valuable self-driving cars are. Self-driving cars are getting well within reach if we were to put the effort into it. An “Apollo” project could provide them in a decade, I’m pretty comfortable. And while they might not be perfect, and that raises its own questions, the 230 billion annual saving is quite a huge payoff.
(And that’s not counting a possibly superior energy payoff as people start using “right sized” vehicles that differ for every trip, owning only the type of vehicle they used most often if they choose, and taking autotaxis — which may be other people’s cars rented out when not in use) for other styles of trips. Such a regime means the vast majority of trips would be in small, fast 1 or 2 person electric city cars/trikes, medium haul trips in the right sized liquid fuel car or long-range electric, and long-range trips in vehicles meant for these, or as combos of short-range-vehicle zips to train and is waiting at next train stop.)
And also let me add that a million die worldwide in car crashes. Lives and delays aren’t perhaps valued with as much money as an NHTSA study would do in the USA, but if they were, that puts the global cost of accidents at 5.3 trillion dollars, and we are blind to it because we accept our cars and their cost so much.
5.3 trillion PER YEAR. Can you think of any problem that while hard, seems very doable from an engineering standpoint, with a payoff like that?
What’s interesting is that, with most people living and working in or near an urban area, almost anyone living in this country not directly involved in agriculture or something similar could easily organize their life in such a way that they do not need to own or drive a car. Surprisingly, given the costs, few do. An economist might make a case that this means that cars deliver a tremendous amount of value to the people that own and use them.
I wonder what the bicycle deaths and related costs (and health benefits) are…
If you included the health effects of having people walk (eg NYC) vs. suburban sedentary life style I’m very sure that would catapult you over the 4% mark.
You haven’t even included all the externalities of costs like pollution (or related deaths) and deaths related to the collection of oil that the cars use – (Northern Iraq comes a bit to mind).
Apparently ethanol is worse than oil as far as pollution. And the dealths are suprisingly high – according to wikipedia just some changes to Diesel standards in the US would reduce dealths by 37,000 per year.
http://en.wikipedia.org/wiki/Air_pollution#Health_effects
In my (purely antectodal) experience, per mile exposure, bicycles are much more dangerous than cars. I ride to work everyday until there is too much snow. In my work I see many injured cyclists, seemingly out of proportion to the number of bikes on the road…
It is amazing to me how the nightly news overlooks the obvious. We keep seeing political debates about health care in America. Health care costs would go significantly down if we cut down on traffic accidents. Fewer people would need health care, and so insurance premiums would go down. Usually, network news stories about cars relate to the cost of oil, ethanol, or pollution. There are 43,000 traffic fatalities annually in this country. There are over 6 million traffic accidents annually in the U.S. There are several engineering techniques which could bring these numbers down:
–A global speed controller. My car speedometer goes up to 140 mph. Why, I have no idea. Is there any other product which you can just press a go button until it kills everyone? Would they sell an oven which, if you turned it to the maximum heat, would heat up until it burned the house down? Why sell a car which will accelerate until you are going at dangerous and illegal speeds?
–A specific speed controller. With GPS technologies, the car can know the speed limit for the road it is on. You could set this to only drive 0, 5, or 10 mph faster than the limit.
–Radar-based technologies to prevent the car from smashing into objects in front of it. These are technologically feasible, but don’t seem to be available on basic models of cars.
The main argument against these is that some people wouldn’t want to buy them. I would argue a few points. First, the car manufacturers have to make them available. You simply don’t know what the market would be unless you offer them Second, they should be available on basic cars. These safety features shouldn’t only be available to people who have $40 K to spend on a car. If some of these features cost $5 K, you should be able to add them to an inexpensive car, and leave with a $15 K purchase, not a $45 K purchase.
Many people would object to the speed controllers on the ground that they are good drivers and only speed in a safe manner. However, lots of people would buy these technologies for their children. (I bet almost everyone would buy these technologies for their children if they were offered at a reasonable price). If these technologies are very helpful in reducing accidents, then the cost of purchasing them would be offset by lower car insurance rates. Moreover, many accident costs are born by the government (police, ambulance, courts, motorists without auto or health insurance). Therefore, the government should give a tax credit to reward people who purchase these technologies. The technologies would probably pay for themselves. So why don’t we see stories about 43,000 deaths on the network news? It is a mark of our collective acceptance of the car culture. There is the cowboy myth, and the car is our modern trusty steed. The freedom that we feel is provided by the car, which some might say we embrace with an almost religious fervor, is described in Bruce Springsteen’s Thunder Road–
“Waste your summer praying in vain
For a saviour to rise from these streets
…All the redemption I can offer, girl, is beneath this dirty hood
…Heaven’s waiting on down the tracks
…We’re riding out tonight to case the promised land”
Unfortunately, we don’t usually get that sense of freedom. What slows us down is not the 140 MPH upper speed limit on the speedometer, but the fact that there’s an accident ahead of us, delaying us for an hour. And almost everyone has had someone in their family in a serious accident–possibly fatal.
responding to 2nd part of philip’s query: cost per mile.
i divide
$520B (4% GDP, from above posting), by
2,347B vehicle miles (from http://www.nhtsa.dot.gov/people/outreach/employer/WhatCost/appen.html for 1994) giving:
$0.22 per mile.
this is about half of the standard 2006 employer reimbursement rate, FWIW. (see http://en.wikipedia.org/wiki/Business_Mileage_Reimbursement_Rate), which likely exceeds the largest single cost/mile component.