A Data-Driven Look at the Rise in Cycling Fatalities
60-something cyclists, riding at night, and the increase in cyclist deaths between 2010 and 2021
More than 900 cyclists were killed in traffic crashes in the US in 2021, an increase of more than 50% since 20101. In this post, I look at when this is happening, what kinds of cars are involved, where it’s happening, and which cyclists are getting hit and killed.
The danger to cyclists is not just related to an increase in miles driven, or an increase in overall fatalities on the roads. In 2007, cyclists made up 1.76% of all road deaths. In 2020, that percentage had increased to 2.46% (the percentage fell slightly in 2021).
In other words, more cyclists are getting killed. As a cyclist and someone interested in public health, I’m eager to figure out what’s driving that.
In this post, I look at four different questions:
Are more cyclists killed during the day or at night?
Are bigger cars and trucks causing the increase in cyclist fatalities?
Where are cyclists getting killed?
Are old cyclists dying more, or is it young ones?
There are some small changes in where and when cyclists are getting killed, and some very large changes in who is getting killed.
The vast majority of cyclists killed on the road are male: 87% of cyclists killed since 2007 have been male, and that percentage was steady from 2007 through 2021.
Are more cyclists killed during the day or at night?
From 2007 through 2014, there were more cyclists killed during daytime hours than at night. But the number of nighttime cyclist fatalities has been increasing more quickly than the number of daytime cyclist deaths, and more cyclists are now killed at night. In 2021, 52% of cyclist fatalities were at night, compared to 44% in the day.2
That means that the vast majority of the increase in cyclist deaths has been at night: 195 more cyclists were killed at night in 2021 than in 2007, while 54 more cyclists were killed during the day in 2021.
The increased risk is similar to the trend we see with pedestrian fatalities.
The roads have consistently been more dangerous for pedestrians at night than during the day, but the increase in risk to pedestrians since 2010 is almost entirely from higher fatality numbers at night.
The above graph looks at four different types of traffic deaths, by time of day, from 2017 to 2021. Pedestrians are hugely disproportionately killed at night — over half of pedestrian deaths are between 6 PM and midnight. Cyclists are killed disproportionately at night, but not as disproportionately as pedestrians are. Drivers and passengers deaths are relatively evenly split across the hours of the day.
These patterns align with the level of visibility at night: pedestrians are generally the least visible at night, cars are the most visible, and cyclists fall somewhere in between.
Are bigger cars and trucks causing the increase in cyclist fatalities?
It does not appear that vehicles getting larger is a major factor behind the increase in bike fatalities.
There is an increase in cyclists killed by both smaller vehicles and larger ones, and the increase is only slightly bigger among large vehicles.
However, large vehicles kill a disproportionate number of cyclists and pedestrians. Over half of cyclists killed and over half of pedestrians killed were struck by larger vehicles (SUVs, light trucks, buses, etc.) — while less than half of primary vehicles3 in fatalities of drivers and passengers were larger vehicles.
Where are cyclists getting killed?
Cyclists — like pedestrians — are disproportionately killed in blue and purple counties. Drivers and passengers are disproportionately killed in red counties.4
There is a similar but less pronounced trend with blue, purple, and red states. Between 2017 and 2021, the states with the most cyclist fatalities were, in order, Florida, California5, Texas, New York, and Arizona.
Cyclists — similar to pedestrians — are also disproportionately killed in the lowest income ZIP6 codes. Drivers and passengers are slightly more likely to be killed in the lowest income ZIPs, but the differences by ZIP code are far smaller.
Cyclists and pedestrians are also killed disproportionately in the counties with the lowest levels of education7.
Are old cyclists dying more, or is it young ones?
This was the biggest surprise finding: there are far, far more old cyclists getting killed today than there were 15 years ago.
Cyclists who were killed on the road between 2017 and 2021 were older than any other group — older than the pedestrians who were killed, older than the drivers who were killed, and older than the passengers who were killed.
And the age mix of cyclist fatalities has shifted quite a bit since 2007. In 2007, 16.3% of cyclists killed were 60 or older and 13.4% were under 18.
Fourteen years later, in 2021, 35% of cyclists killed were 60 or older, and just 6.7% were under 18.
In fact, virtually all of the increase in cyclist fatalities since 2007 has been among cyclists sixty or older:
From 2007 to 2021, there was a 7% increase in deaths of cyclists under 60. Over the same period, there was a 196% increase in deaths of cyclists 60 or older.
It’s also worth noting that the trend of more cyclists being killed at night and the trend of more older cyclists being killed are totally separate: about 70% of cyclists age 60+ are killed during daytime hours, while most younger cyclists are killed at night.
A positive note on road safety
The per capita motor vehicle fatality rate has mostly been trending in the right direction for many decades
Even as cars have gotten faster and people have driven more miles, roads are mostly getting safer. The technology to make vehicles safer, the data to understand what’s going on, and the political and cultural tools for improving infrastructure and culture all exist in the US.
That said, there are multiple negative trends affecting (and killing) both cyclists and pedestrians, and I’ll keep trying to shed light on why. We can do better.
Analysis includes 2007-2021 FARS data from the NHTSA augmented with ZIP code level tax return data from the IRS, CDC data on drug overdoses and general health trends by county, MIT Election Data + Science Lab data on election results by county, and ACS five year data on education and demographics by county.
4% were dawn or dusk or some version of “other.”
I compare the first vehicle listed in multi-vehicle crashes, to have a clear comparison point with crashes involving a single motor vehicle and either a pedestrian or a cyclist.
For this analysis, I defined red counties as counties where an average of at least 57% of the two-party vote share was for Republicans in the 2012, 2016, and 2024 elections. Blue counties were those where at least 57% of the vote share was for Democrats, and purple counties were those where neither party received a 57% share.
I’ll note two things here: one, warmer states tend to have more cyclist fatalities. Second, Florida’s per capita cyclist death rate is about double California’s.
Technically this analysis was done with ZCTAs.
This analysis was done based on education levels for females; results based on education levels for males is almost identical.
Interesting to see the similarities between pedestrian and bicycle deaths. I believe pedestrian deaths largely occur on "non-highway arterials." Is that data available for bicycle deaths? Based on victim demographics and location, I wonder how much poverty and homelessness could be to blame for both.
Very interesting. As a relatively new rider (7 years) who rides with a group composed of mainly over 50 y/o riders, I wonder if the 60+ fatalities is related to an increase in 60+ cyclists? Would be interesting to determine if drivers' cell phone usage while driving has gotten worse during the test period? I'm guessing it has