« Why Not? | Main | What I've Been Reading »
Detroit fact of the day
Among cities with more than 500,000 residents, Detroit has the safest drivers, with accident rates that are 20 percent below the national average.
For cities with more than 1 million residents, Phoenix has the safest big-city commuters, with accident rates about equal to the national average.
Here is much more, Philadelphia is a disaster and L.A. isn't so safe either. I wonder to what extent these rankings simply proxy for traffic density. Here are some charts. Overall Sioux Falls, Tucson, and El Paso seem to be relatively safe cities for driving.
Posted by Tyler Cowen on July 3, 2008 at 03:37 PM in Data Source | Permalink
Comments
As a native Detroiter who frequently returns home (I was just there a few days ago), I should note that this stat is even more impressive given that Detroiters drive VERY fast. Even on urban freeways, you better be doing 70 or 75 or you're going to get passed left and right.
Posted by: Steve Horwitz at Jul 3, 2008 3:56:01 PM
Data on metropolitan areas would be more valuable than data on cities, as is almost always the case.
Posted by: at Jul 3, 2008 4:16:15 PM
Data on metropolitan areas would be more valuable than data on cities, as is almost always the case.
Posted by: at Jul 3, 2008 4:16:39 PM
These numbers don't make any sense. Sioux Falls, SD ranks as the safest city with an average of 14.6 years between accidents. But it doesn't appear on the top ten for males or females. So males must have an average below 8.4 years between accidents, and females below 8.2 years between accidents. Is there a large population of very safe ungendered drivers in Sioux Falls? Or am I missing something obvious?
Posted by: wintermute at Jul 3, 2008 4:29:11 PM
Based on our frustrating experience of getting timely crime data out of Detroit, I'd question the reporting. It might be a good place to have an accident go unreported.
Posted by: Colin at Jul 3, 2008 4:32:09 PM
Most traffic in southeastern michigan isn't in Detroit. It's on the expressways that run all around it. The data is most likely already skewed.
Posted by: John at Jul 3, 2008 4:38:18 PM
The most dangerous roads, I recall reading somewhere, are "local arteries": streets with lots of intersections that carry lots of through traffic at 35+ Mph. These streets are distributed very unevenly; certainly Philadelphia is dominated by them and LA, famous for its freeways, may have a lot too.
Posted by: Sam Penrose at Jul 3, 2008 4:42:36 PM
Los Angeles' freeways are modern day deathtraps. If quality of life wasn't relatively high, I would not live here.
I only drive about 10-15 minutes to get to work from home, but even in that short time I almost always fear for my life at some point. The main problems that I can tell: a) most people are selfish drivers and b) there is little to no accountability for poor driving. Police officers are too easily spotted and avoided.
Posted by: Brian at Jul 3, 2008 5:12:57 PM
LA Freeways? I almost died on Sunday when some nut half-ran a red light on San Vincente at Wilshire. Once he realized he was going to hit the car next to me he slammed on his brakes, the car next to me swerved into my lane, I swerved into the (empty) oncoming lane.
The other problem in LA is that if you accidently cut someone off, you might get shot.
Posted by: Mr. Econotarian at Jul 3, 2008 6:23:52 PM
FYI, the speed limit on many highways in Detroit is 70 mph...does anyone know how the city does versus others when looking at highway accidents?
Posted by: Samir Nurmohamed at Jul 3, 2008 6:24:09 PM
Accident rates and safety are NOT the same thing.
For example, our car was recently rear-ended here in Denver. At 5 mph. An accident? Yes. A safety risk? Emphatically not.
Posted by: Dave Barnes at Jul 3, 2008 6:47:47 PM
That there is actually a difference between men and women is actually kind of suprising to me. Sterotypes aside, you would think that a woman would be more likely to have a child in the car and (hopefully) be more cautious.
Also suprising that the biggest gap between men and women is in the youngest age group.
I wonder how this data compares to pre-cell phone era data?
Posted by: JohnZ at Jul 3, 2008 6:56:57 PM
JohnZ, I think the possibility that women are more likely to have kids in the car would cause them to have more distractions. Try having two screaming, fighting or otherwise children in the back seat of your car and I'm willing to bet you become frustrated, angry, or any other number of emotions that could be very distracting.
my $.02.
Posted by: TomM at Jul 3, 2008 7:07:53 PM
I wish someone would compile these stats in Australia. You could do some nice graphics with these tables.
Posted by: Kris at Jul 3, 2008 8:36:38 PM
Phoenix factors are probably low population density, wide roads, straight grid-patterned layout and good year round weather conditions.
Posted by: tom.a at Jul 3, 2008 10:02:34 PM
Since these are based on insurance claims, couldn't a high fraud city like Philadelphia "show" more accidents than others?
Posted by: CD at Jul 4, 2008 3:06:44 AM
CD - Philadelphia may have a high fraud rate, but it also has the highest (or one of the highest) rates of uninsured motorists too. I doubt that accidents between two uninsured motorists get reported often: why bother?
Posted by: liberalarts at Jul 4, 2008 8:32:11 AM
CD - Philadelphia may have a high fraud rate, but it also has the highest (or one of the highest) rates of uninsured motorists too. I doubt that accidents between two uninsured motorists get reported often: why bother?
Posted by: liberalarts at Jul 4, 2008 8:32:24 AM
CD - Philadelphia may have a high fraud rate, but it also has the highest (or one of the highest) rates of uninsured motorists too. I doubt that accidents between two uninsured motorists get reported often: why bother?
Posted by: liberalarts at Jul 4, 2008 8:32:28 AM
Those of us in Michigan have some questions about that study. Notably, how are they counting those crashes? If the insurance company is looking at insurance claims, that would be an issue. There are parts of the city where the majority of vehicles are reportedly uninsured [citation needed]. If they use police reports, would you call the police if you were an uninsured driver in a crash? I reviewed crash forms from the Detroit intersection with the most crashes, and quite a few fender benders were hit and runs where the "running" driver stopped, made sure no one was injured, then left before the police came (saying they had no insurance/license/whatever). Few of those minor crashes where two uninsured parties collide are likely to be reported.
Posted by: Zubon at Jul 4, 2008 8:44:00 AM
Tyler's point about congestion is spot on. Detroit's economy is awful, and as a result no one is on the road. My parents live there, and when they visit me in Houston they always are blown away by how many people are on the freeways, saying that our weekend traffic is more dense than some Detroit freeways in rush hour.
Not to mention that with so many folks working for the auto companies, in the Detroit area, a given household owns more cars than one with comparable wealth in another metro, because the cars are cheaper with employee discounts. And "number of cars" (rather than the more appropriate lane-miles) is the denominator, further biasing the results for Southeast Michigan (note the various other Detroit-area cities on the list, too.
Posted by: bottomofthe9th at Jul 4, 2008 8:59:05 AM
The measure -- years between accidents -- does not seem to be corrected for miles driven or time spent in one's car. In cities like Los Angeles, where drivers typically spend more of their life in their automobile, one would expect years between accidents to be smaller, even if accidents per mile were lower than in some other cities.
Posted by: anonymous at Jul 4, 2008 11:15:35 AM
bottomofthe9th, Steve Horwitz:
While there may be statistical anomalies that makes Detroit look better than it really is, Detroit drivers really are much better than drivers in most other places.
My theory is this:
1. In Michigan, many people take drivers education as a part of their high school education. I took drivers education and auto safety in high school, for an entire semester, for a real grade. My high school at the time actually had a dedicated auto-track to drive on, we did obstacle courses through cones and barriers, etc.
2. People don't engage in as much aggressive driving, because they know better than to mess with random strangers.
3. Less youth street racing.
Those of us in Michigan have some questions about that study. Notably, how are they counting those crashes? If the insurance company is looking at insurance claims, that would be an issue. There are parts of the city where the majority of vehicles are reportedly uninsured [citation needed]. If they use police reports, would you call the police if you were an uninsured driver in a crash? I reviewed crash forms from the Detroit intersection with the most crashes, and quite a few fender benders were hit and runs where the "running" driver stopped, made sure no one was injured, then left before the police came (saying they had no insurance/license/whatever). Few of those minor crashes where two uninsured parties collide are likely to be reported.
Zubon: White Suburbanite Urban myth. Poor people, who would be the type of people who don't have insurance, are also the type of people who are also of the demographic (junky cars, dark skin) that cops especially like to pull over and harass. A black man with a junky car is absolutely the last person in America who can get away with driving around with no insurance.
Posted by: Rex Rhino at Jul 4, 2008 11:38:55 AM
White Suburbanite Urban myth [citation needed] That's a significant claim about Detroit PD, especially given the similar demographics of officers (and minority leadership of every agency with jurisdiction in the city). Two claims really: racial profiling and that police in Detroit have nothing better to do than harass drivers.
Seriously, I am not making up the narratives on the crash forms. You can check them yourself these days. We work with Detroit PD on traffic enforcement, so I don't think I have a completely unrealistic view of their priorities in pulling people over.
Posted by: Zubon at Jul 4, 2008 1:53:32 PM
Both traffic density and type/date of road construction are important, I'm sure that once you remove these two factors there's no longer any meaning to the inter-city data.
I'm surprised how high the overall frequencies are! If I count the family I grew up in, where I have complete data, it's 5 accidents in 107 years of driving, ages 15-58 (and only 2 in 68 years if you leave out my mom).
Posted by: Paul N at Jul 4, 2008 2:47:49 PM
