Search for motorcycle commuting and the first page of Google hands you the same article ten times: ten tips, be visible, check the weather, pack spare gloves. It is written by a dealership in Florida, a dealership in California, an accessory shop, a riding school. Above all of them sits a Reddit thread asking the only question anyone actually types into the box — how dangerous is this? — and below them sits a Facebook group arguing about it.
I commute on a motorcycle, I build an app for people who ride them, and none of those pages told me anything I could check. So I went looking for the documents instead: the Census table that counts how many Americans actually do this, the federal crash report that says where the risk really lives, the six state laws that decide what you may do in stopped traffic — two of which, I found out, are written to expire — and the university study those laws were built on.
What the documents say is more interesting than the tips. It is also, in places, the opposite of what the first page of Google assumes.
The short version, before the documents
- The Census Bureau’s 2023 survey counts 215,105 Americans whose usual way to work is a motorcycle — about one worker in 755. More than three times as many commute by bicycle.
- A modest commute makes you a statistical outlier in the other direction too: the average American motorcycle covers 2,121 miles a year, and ten miles each way to work is roughly double that on its own.
- Per mile travelled, riding is about 28 times deadlier than sitting in a car — but the fatal-crash ledger is heavily weighted toward nights, weekends and alcohol, which is to say: away from the sober morning ride to work. One line in the licensing manual, though, points straight at commuting, and it is the five-mile line.
- Six states now allow some form of riding between or past other vehicles. Five of them wrote numbers into the law — 15, 20, 25 miles per hour — and those numbers trace back to one study. Only California allows it in moving traffic, and its law contains no numbers at all.
- Two of the six laws have expiration dates. Colorado’s repeals itself on September 1, 2027, and its transportation department owes the legislature a safety report by January 1, 2027. Whether filtering stays legal there is, quite literally, an experiment with a lab report pending.
Every one of those claims has a document behind it. Here they are.
The rarest way to get to work in America
The American Community Survey asks workers how they usually got to work in the previous week; if someone used more than one mode, they answer with the one covering the most distance. Table B08301 breaks the answers out, and it has a line item most people never notice: Motorcycle.
For 2023, the numbers for the United States look like this: 162,434,675 workers. 126,985,709 of them usually drove or rode in a car, truck or van. 22,486,510 worked from home. 3,966,159 walked. 761,757 rode a bicycle.
Motorcycle: 215,105 — give or take 9,291, since this is a survey estimate with a published margin of error.
Divide it out — the division is mine, the numbers are the Census Bureau’s — and one American worker in 755 commutes by motorcycle. Cyclists outnumber motorcyclists three and a half to one. Set against the Federal Highway Administration’s count of 9,516,910 registered motorcycles, the same arithmetic says that very roughly one registered bike in forty-four is anyone’s usual way to get to work. Those are two different kinds of numbers — a stock of machines against a count of people — so treat that last one as an order of magnitude, not a statistic. The order of magnitude is the point: in America, the commuter is the exception among riders, and the rider is a rounding error among commuters.
That rarity is worth taking seriously, because it explains things the tips articles treat as bad luck. Drivers do not expect you, because they can go weeks without seeing one of you at rush hour. Parking rules, employer policies and road design were written for the 127 million, not the 215 thousand. The federal safety report I will get to shortly puts the driver side of this plainly: a motorcycle’s smaller size, it warns drivers, “can be hidden in your vehicle’s blind spot”. For what it is worth, the average one-way commute across all modes was 26.8 minutes in 2023 — the Census publishes that too.
What a commute does to your odometer
The Federal Highway Administration’s Table VM-1 puts the average distance travelled per motorcycle in the United States at 2,121 miles for 2023. I went through that number in detail in the mileage article, including the years around it — the series since 2018 never leaves the corridor between roughly 2,100 and 2,600 miles.
Now do the commuter’s arithmetic, which is mine, not the FHWA’s. Ten miles each way, five days a week, allowing for holidays and the days you take the car — call it 240 riding days: 4,800 miles a year. More than double the national average, before a single Sunday ride. Even a short five-mile hop each way clears the average on its own.
This is not a trivia point. It flips the single most useful rule in motorcycle maintenance. When I read the factory schedules for the maintenance schedule article, the conclusion that fell out of the manufacturers’ own numbers was that at average annual mileages, the calendar items — fluids by years, rubber by age — come due long before the mileage columns do, so for most riders the odometer is the secondary trigger. The commuter is the reader for whom that sentence runs backwards. At 4,800 miles a year — my arithmetic from a moment ago — a 4,000-mile chain service arrives every ten months; at scooter-commuter distances it arrives faster still. You are the rider the mileage columns were printed for.
The wear pattern is different in kind, not just in speed. A chain doing twelve kilometres twice a day in all weather lives a completely different life from one doing two hundred on dry Sundays — the chain guide goes into why one maker counts rides rather than distance, and a commuter racks up rides at a rate no touring rider can match. Cables and pivots on a bike parked outside an office in the rain want attention on their own rhythm, which is a thing I learned writing about clutch adjustment. Tyres are the one place commuting is kind to you: the harder compounds that touring and commuting call for are the long-lived end of the tyre guide‘s spectrum. And the battery, which hates short runs and stop-start traffic, gets its own winter chapter in the battery article.
None of that is a reason not to commute. It is a reason to know your real annual mileage — which almost nobody does, and which I will come back to.
How dangerous is it, actually
This is the question the Reddit thread at the top of the search results is really asking, so I will answer it with the document that exists to answer it: NHTSA’s Traffic Safety Facts for motorcycles, 2023 data, published in July 2025 as report DOT HS 813 732.
The headline number is brutal and worth quoting exactly:
Per 100 million vehicle miles traveled in 2023, the fatality rate for motorcyclists (31.39) was almost 28 times the passenger car occupant fatality rate (1.13).
Twenty-eight times, per mile travelled. The same report gives the gentler-looking version — per 100,000 registered vehicles the ratio is about six to one — but the per-mile figure is the honest one for a commuter, because a commute is measured in miles, not in registrations. 6,335 riders died in 2023, 15 percent of all traffic deaths, from machines that were 3.1 percent of registered vehicles and 0.6 percent of miles driven.
So the answer to “is it dangerous” is yes, by any measure. But the report also describes the shape of the danger, and the shape is not commute-shaped.
Start with the clock and the calendar. NHTSA defines the weekend as Friday 6 p.m. to Monday 5:59 a.m. — two and a half days — and reports that 47 percent of riders killed in 2023 died inside that window. Spread the other 3,371 weekday deaths over the remaining four and a half days and you get roughly 749 per weekday-day against roughly 1,186 per weekend-day. That division is mine, and it comes with an honest caveat: NHTSA does not publish how many miles are ridden in each window, so this is a count, not a rate — if weekend riding miles dwarf weekday miles, the picture changes. But the raw ledger leans hard toward the weekend.
Then the alcohol columns, which are the report’s most repeated finding:
Forty-one percent of motorcycle riders who died in single-vehicle crashes in 2023 were alcohol-impaired.
Motorcycle riders killed in traffic crashes at night were two and a half times more frequently alcohol-impaired than those killed during the day (38% and 15%) in 2023.
A sober rider on a weekday morning is simply not the protagonist of those statistics. I want to be careful here — the risk multiple per mile applies to everyone, and nothing in this section says commuting is safe. What the ledger says is narrower and more useful: a large share of the fatal-crash total is generated by circumstances a commuter can simply decline to be in — night, alcohol, weekend recreational speed. Ride sober at 8 a.m. and you have already opted out of the single most lethal pattern in the book.
What you cannot opt out of is the crash type that belongs to exactly the environment a commute runs through:
In 2023 there were 3,419 fatal two-vehicle crashes each involving a motorcycle and another type of vehicle. In 46 percent (1,588) of these crashes, the other vehicles were turning left while the motorcycles were going straight, passing, or overtaking other vehicles.
The car turning left across your path, at an intersection, in traffic — that is the commuter’s crash. The same report notes that in two-vehicle fatal crashes, 79 percent of the motorcycles were struck on the front. Only 6 percent were hit from the rear, a number worth remembering when the filtering laws come up below. The licensing manual condenses the geography of all this into one sentence I have quoted before and will keep quoting: “The greatest potential for conflict between you and other traffic is at intersections.” The habits that defend against exactly this — lane position, cover the brakes, assume the left-turner has not seen you — are the whole subject of the defensive riding guide, and the braking system that saves the panicked version of that moment has its own article.
The five-mile line in the license manual
There is one sentence in the Motorcycle Safety Foundation’s operator manual — the booklet American licensing exams are built on — that is about commuting without ever using the word:
Most crashes happen on short trips (less than five miles long), just a few minutes after starting out.
Read that against your own ride to work. A commute is the short trip, repeated five hundred times a year. The manual deploys that fact in an argument about helmets, and the two sentences that follow it complete the thought:
Most riders are riding slower than 30 mph when a crash occurs. At these speeds, helmets can cut both the number and the severity of head and neck injuries by half.
Note what this does to the most common commuting rationalisation there is — it is only ten minutes, it is only around the corner, I will just slip through town. The trip you are tempted to under-dress for is statistically the trip the crashes happen on, at speeds where the helmet does its best work. That halving claim is specifically about head and neck injuries at those crash speeds, not a promise about surviving anything; the wider case for what you wear on a daily basis is in the gear guide, which is honest about the fact that commuter gear is its own category with its own compromises, and the question of when a daily-wear helmet is finished has an article of its own — spoiler: the maker’s clock runs faster for exactly the person who rides twice a day, every day.
The rest of the daily discipline I will only point at, because each piece has its own deep dive: the five-minute pre-ride check matters more when the bike sleeps outside an office; rain and darkness are not occasional conditions for a commuter, they are November; and a bike parked in the same public spot for eight hours a day, five days a week, is a theft target with a schedule — the workplace-parking problem gets its own section in the anti-theft guide.
Filtering and splitting: six states, five rulebooks, two expiry dates
Here is where the documents get genuinely surprising, because almost everything the internet repeats about lane splitting in America is either too generous or too vague.
Start with the only state everyone names. California’s entire statute on the subject, Vehicle Code section 21658.1, does two things. First, it defines the practice:
“lane splitting” means driving a motorcycle, as defined in Section 400, that has two wheels in contact with the ground, between rows of stopped or moving vehicles in the same lane, including on both divided and undivided streets, roads, or highways.
Second, it authorises the Highway Patrol to publish educational guidelines about it. That is all. The California law contains no speed limit, no traffic-condition test, no numbers of any kind — it recognises the practice, including between moving vehicles, and delegates the wisdom to the CHP. It took effect on January 1, 2017.
Every other state that has legalised anything has legalised something narrower — filtering, not splitting — and each wrote its own numbers. I pulled all five statutes; here is what they actually require, condensed:
| State | In force since | You may pass | Your max speed | Key conditions |
|---|---|---|---|---|
| California | Jan 1, 2017 | Stopped or moving traffic | No number in the statute | CHP guidelines, not statutory limits |
| Utah | 2019 (sunset now Jul 1, 2027) | Stopped vehicles only | 15 mph | Speed limit ≤45 mph, ≥2 lanes same direction, off-ramps yes, on-ramps no |
| Montana | 2021 | Stopped or ≤10 mph vehicles | 20 mph | Lanes wide enough to pass safely |
| Arizona | 2022 | Stopped vehicles only | 15 mph | Speed limit ≤45 mph, ≥2 adjacent lanes same direction |
| Colorado | Aug 7, 2024 (repeal Sep 1, 2027) | Stopped vehicles only | 15 mph | Adjacent lanes stopped too; must stop filtering when traffic moves; never on the right shoulder |
| Minnesota | Jul 1, 2025 | Slower traffic | 25 mph, and ≤15 mph over traffic speed | Not in roundabouts, school zones, single-lane work zones, queued on-ramps |
The details differ in ways that would matter enormously on an actual Monday. In Utah the vehicle you pass must be stopped, full stop — the moment the queue creeps, the legal basis evaporates; the statute even distinguishes ramps, permitting filtering on off-ramps and forbidding it on on-ramps. Montana lets you pass traffic that is still rolling at up to 10 miles an hour, and gives you 20. Colorado is the strictest of the set: the vehicles beside you must be stopped as well as the one you are passing, and the act says in terms — in the capital letters American legislatures use for newly enacted text — that when the queue starts moving, “THE DRIVER OF THE MOTORCYCLE SHALL CEASE OVERTAKING OR PASSING”. Minnesota, the newest, is the only filtering state that lets you pass moving traffic: not more than 25 miles per hour, and not more than 15 over the speed of the traffic you are passing.
Everywhere else, the default rule applies, and Arizona’s own code states the default with unusual clarity before carving its exception: the operator of a motorcycle shall not overtake and pass in the same lane occupied by the vehicle being overtaken, and shall not ride between lanes of traffic or between adjacent rows of vehicles. In most of the remaining forty-four states that pair of sentences, in some local wording, is still the whole law. Not authorised is the accurate phrase — a few states have no statute naming the practice at all, and what is not authorised gets ticketed under lane-usage or careless-driving rules.
Now the detail that genuinely changed how I think about this. Two of the six laws are written to expire. Utah’s filtering provision carried a sunset from the day it passed in 2019; the 2022 legislature, after a review, extended it to July 1, 2027. Colorado’s is more dramatic — the subsection that permits filtering announces its own end date:
THIS SUBSECTION (3)(b) IS REPEALED, EFFECTIVE SEPTEMBER 1, 2027.
And the same act orders the homework: the state Department of Transportation must analyse rear-end collisions and sideswipes before and after the law took effect and report to the legislature by January 1, 2027. American lane-filtering law is not a settled question — it is a running experiment, with a lab report due.
Where the 15 mph everywhere comes from
Look back at that table: 15 appears in four of the six rows. That number has a specific origin, and it is not a legislature’s intuition.
In 2015, researchers at UC Berkeley’s Safe Transportation Research and Education Center — supported by NHTSA through a grant from California’s Office of Traffic Safety — analysed 5,969 motorcycle collisions collected by police across California over fourteen months, including 997 involving riders who were splitting lanes when they crashed. Their conclusion is the closest thing this subject has to a founding document:
Lane-splitting appears to be a relatively safe motorcycle riding strategy if done in traffic moving at 50 MPH or less and if motorcyclists do not exceed the speed of other vehicles by more than 15 MPH.
The study found that injury rates did not rise with speed differentials up to 15 mph over traffic; above that, every injury type climbed. It also produced the statistic that built the safety case for filtering: lane-splitting riders were “significantly less likely to be rear-ended” than riders who stayed in the queue — 2.6 percent against 4.6 percent. Remember NHTSA’s geometry from earlier: 79 percent of motorcycles in fatal two-vehicle crashes are hit on the front, 6 percent on the rear — but the rear-end crash in stopped traffic is precisely the one where the car does not brake and the rider has nowhere to go. The study is equally honest in the other direction, which is why I trust it: the same lane-splitting riders were considerably more likely to rear-end somebody else — 38 percent of their crashes against 16 percent. Filtering trades away the crash you cannot prevent and buys more of the one you can.
When Colorado’s legislature wrote its 2024 law, it stacked the evidence into the bill itself. The legislative declaration cites a 2009 five-country European in-depth study finding that 0.45 percent of motorcycle crashes involved lane splitting and that riders were seven times more likely to be hit while stopped than while splitting; two studies reporting lane-splitting riders “43 percent less likely to be involved in a rear-end crash”; and a claimed 30 percent drop in motorcyclist fatalities in the year after the CHP published its safety tips. Those are the legislature’s findings, quoted from the act — I have not verified them independently, and the causal reading of that last one is theirs. But they show what convinced six statehouses, and they are why the numbers in the table are the numbers they are: the laws are, more or less, the Berkeley envelope with a safety margin shaved off.
If you filter where it is legal, the envelope is the practical takeaway: past stopped or crawling traffic, at a walking-pace differential, positioned where drivers’ mirrors and hand signals can still do their work. If you ride where it is not — that is most of America — the rear-end statistic is instead an argument for obsessive mirror-watching at the back of every queue.
The green argument, audited
The last standard argument for commuting by motorcycle is environmental: smaller machine, less fuel, greener commute. The documents split this one down the middle, and the honest version is worth having.
On fuel, the bike wins and it is not close. The FHWA table that gives motorcycles 2,121 miles a year also carries a fuel-economy column: 44.0 miles per gallon, against the mid-twenties for light-duty vehicles. One caveat I flagged in the mileage article and will repeat: that 44.0 has been identical in every edition of the table I have checked, from 2019 through 2023, while the car figures move — it is a modelling coefficient, not a measurement. Directionally it is obviously right; a 200-kilogram machine moves one body on a fraction of the fuel a two-tonne one does, and burning less fuel is emitting less carbon dioxide.
On everything else that comes out of an exhaust, the ledger reverses, and you can read the reversal directly in the Code of Federal Regulations. The federal emission standard for a Class III motorcycle — 280 cc and over, which covers most full-size machines; the smaller classes that commuter scooters fall into have looser rules still (where that boundary actually sits, and why a 49 cc machine can be a moped in one state and a motorcycle in the next, is in moped vs scooter) — has been the same since it phased in for the 2010 model year: hydrocarbons plus nitrogen oxides, 0.8 grams per kilometre; carbon monoxide, 12.0. Sixteen years, no tightening. Cars, meanwhile, live under Tier 3, whose fully phased-in fleet average for the equivalent smog-forming basket is 0.030 grams per mile.
Convert the units and divide — the arithmetic is mine, and the two rulebooks count their hydrocarbons in slightly different baskets, so treat this as scale rather than precision: the motorcycle standard allows roughly forty times the smog-forming emissions per mile of the average new car, and between four and five times the carbon monoxide. These are limits on what may be sold, not measurements of what your particular bike emits. But as a statement about how hard the two industries have been pushed, it is unambiguous — the modern car has been regulated into near-cleanliness while the motorcycle standard has stood still since 2010.
So the honest sentence is: a motorcycle commute burns meaningfully less fuel and carbon, and emits meaningfully more smog, than the same commute in a modern car. Anyone who tells you only half of that is selling something. If the whole ledger bothers you, the genuinely clean version of two-wheeled commuting exists and I have written about electric motorcycles — commuting distances are, if anything, the use case where their range limits stop mattering.
What the commute does for you
I have spent four thousand words on documents, so let me spend one paragraph on the part no table captures, because it is the actual reason those 215,105 riders do it.
A commute is the worst version of driving — the same road, twice a day, in traffic. It is somehow the best version of riding. The lane you hold, the gap you read, the roundabout you get right: a motorcycle turns the two dead half-hours of the day into the only part that demands you be entirely present. Every rider who commutes says some version of this; the licensing manual would call it risk management, and it is, but it is also the difference between arriving annoyed and arriving awake. That is an opinion, clearly labelled. Everything else in this article is a citation.
Where RideLog fits, and where it does not
This is the part where I declare an interest, so I will be blunt about it: I build RideLog, a motorcycle app for iPhone and Android, and a commuter is almost the perfect stress test for what it does.
The commuter’s defining data problem is that no single ride feels worth recording — twelve minutes, the same road — so nothing gets recorded, and at the end of the year you have no idea whether you rode two thousand miles or six. I wrote above that the whole maintenance question turns on your real annual mileage; the commuter is the rider least likely to know it and most affected by it. RideLog detects rides automatically from movement and GPS — no start button, no stop button — so the five hundred small trips a year accumulate into the one number that decides whether your chain, your oil and your tyres live on the calendar or on the odometer. It logs fuel stops and works out what the stop-and-go miles actually cost you. It keeps each bike’s history separately if you run a commuter and a weekend machine, which — as the chain and clutch articles spell out — is exactly the case where the two schedules diverge. And the maintenance reminders hold the dates the calendar items come due, next to the distances the mileage items come due.
The privacy design matters more here than anywhere else in the app, and I want to say it precisely. A commute log is, by definition, a record of where you live, where you work, and what time you pass every point between them, five days a week. RideLog keeps all of it on your phone: no account, no cloud, no server of mine holding your routes — there is nothing to breach and nobody to subpoena, because the data never leaves the device. It works offline. When I wrote about sharing rides in online groups, the one rule I insisted on was never to post your home location or your fixed commute times; an app whose architecture makes that mistake impossible is the same rule, built in.
What it does not do, and I would rather say it than have you find out: it is not a navigator and will not route you around traffic. It does not know the filtering law in your state — the table above and your own statute do that. It cannot detect a crash or call anyone. And the distance it records is the distance your phone travelled, from GPS: accurate enough to tell you which maintenance clock governs your bike, not a legal odometer reading. How it compares with the apps that do navigate is an honest question with an honest answer.
What I would tell someone starting Monday
If a friend told me they were about to start commuting on a motorcycle, the documents in this article compress into five sentences.
You are joining a group small enough that nobody on the road is expecting you — one worker in 755 — so ride like you are invisible, because to the left-turning driver in NHTSA’s table, you statistically are. Your risk is real but it is not shaped like your ride: stay sober, favour daylight, respect the intersection, and you have declined most of the ledger. Dress for the five-mile trip, because that is the one the crashes happen on. Learn your state’s actual filtering statute, not the version in a forum comment — the differences between stopped and slow, between 15 and 25, are the difference between a commute and a citation. And know your real annual mileage, because it has quietly doubled, and every service interval on your bike now arrives on the odometer instead of the calendar.
Then, if you want that last number measured instead of guessed, download RideLog free and let a month of commuting count itself. The app was built by someone who does this every working day, which is why the commuter’s problems — the forgettable rides, the two schedules, the route that is nobody’s business — are the ones it solves first.
Frequently asked questions
How many people commute by motorcycle in the United States?
The Census Bureau’s American Community Survey for 2023 estimates 215,105 workers whose usual means of transportation to work was a motorcycle, with a margin of error of about 9,291, out of 162.4 million workers. That is about 0.13 percent — one worker in 755, by my division. For comparison, the same table counts 761,757 bicycle commuters, roughly three and a half times as many.
Is commuting by motorcycle dangerous?
Riding carries a real risk premium: NHTSA’s 2023 data puts the motorcyclist fatality rate per 100 million vehicle miles at 31.39, almost 28 times the passenger-car figure of 1.13. But the fatal-crash ledger is concentrated away from commuting conditions — 47 percent of deaths fall in the Friday-night-to-Monday-morning window NHTSA defines as the weekend, 41 percent of riders killed in single-vehicle crashes were alcohol-impaired, and night-time deaths involve alcohol two and a half times more often than daytime ones. A sober daytime commuter has opted out of the deadliest patterns, though not out of the intersection left-turn crash, which accounted for 46 percent of fatal two-vehicle motorcycle crashes in 2023.
Is lane splitting legal outside California?
True lane splitting — riding between rows of stopped or moving vehicles — is legal in the everyday sense only in California, whose statute defines the practice and sets no speed numbers. Utah, Montana, Arizona, Colorado and Minnesota permit the narrower practice of lane filtering, each with its own conditions: Utah, Arizona and Colorado only past stopped vehicles at up to 15 mph; Montana past vehicles moving at up to 10 mph, at up to 20 mph; Minnesota, since July 1, 2025, at up to 25 mph and no more than 15 mph over the traffic’s speed. Everywhere else the practice is not authorised, and the default rule against passing in the same lane applies.
Why do lane filtering laws expire?
Two of the six laws carry sunset clauses. Utah’s filtering law was born with one in 2019 and, after a legislative review, was extended to July 1, 2027. Colorado’s 2024 law states that the permitting subsection “is repealed, effective September 1, 2027”, and requires the state Department of Transportation to report rear-end and sideswipe collision data to the legislature by January 1, 2027. The legislatures are treating filtering as an experiment: if the safety data disappoints, the laws simply lapse.
Does commuting change how often a motorcycle needs maintenance?
It changes which schedule governs it. The average American motorcycle covers 2,121 miles a year (FHWA, 2023), at which rate the calendar items — brake fluid by years, tyres by age — come due before the mileage columns. A ten-mile-each-way commute roughly doubles that annual mileage by my arithmetic, so the mileage-based intervals for the chain, oil and tyres start arriving first. Daily all-weather use also accelerates chain and cable wear specifically, and short stop-start runs are the pattern batteries tolerate worst.
Is commuting by motorcycle cheaper and greener than driving?
On fuel, yes: the FHWA’s fleet figure for motorcycles is 44.0 miles per gallon — a modelling coefficient rather than a measurement, but directionally right — against the mid-twenties for cars, with proportionally lower carbon dioxide. On smog-forming emissions the comparison reverses: the federal motorcycle standard, unchanged since it phased in for 2010, allows 0.8 g/km of hydrocarbons plus nitrogen oxides, while modern cars certify under a Tier 3 fleet average of 0.030 g/mile — roughly forty times stricter per mile, by my conversion of standards that use slightly different measurement baskets. Cheaper and lower-carbon, yes; cleaner in the smog sense, no.
