Seat Belt Usage and Compensating Behavior

There is a theory that people have an inherent risk thermostat that seeks out an optimal level of risk. When something becomes inherently safer—a law is passed requiring motorcycle riders to wear helmets, for example—people compensate by riding more recklessly. I first read this theory in a 1999 paper by John Adams at the University of Reading, although it seems to have originated with Sam Peltzman.

In any case, this paper presents data that contradicts that thesis:

Abstract—This paper investigates the effects of mandatory seat belt laws on driver behavior and traffic fatalities. Using a unique panel data set on seat belt usage in all U.S. jurisdictions, we analyze how such laws, by influencing seat belt use, affect the incidence of traffic fatalities. Allowing for the endogeneity of seat belt usage, we find that such usage decreases overall traffic fatalities. The magnitude of this effect, however, is significantly smaller than the estimate used by the National Highway Traffic Safety Administration. In addition, we do not find significant support for the compensating-behavior theory, which suggests that seat belt use also has an indirect adverse effect on fatalities by encouraging careless driving. Finally, we identify factors, especially the type of enforcement used, that make seat belt laws more effective in increasing seat belt usage.

Posted on April 11, 2008 at 1:44 PM48 Comments


Jim J April 11, 2008 2:10 PM

I recall something from my graduate economics course – mandatory seat belt laws actually caused the number of pedestrian deaths to slightly increase. As driverrs wore their seat belts and felt incrementally more ‘safe,’ they drove a bit more recklessly. Pedestrians, who got no immediate safety benefit from the seat belt laws, were worse off as a result. I wish I had a citation for this, sorry.

ax0n April 11, 2008 2:31 PM

This likely explains why I always feel naked without my seat belt… and why I always seem to drive like an asshole.

Albatross April 11, 2008 2:51 PM

I suspect that seat belt usage offers several orders of magnitude more risk protection than any increased risk they may cause by recklessness.

I inadvertently did an interesting experiment once. I turned on an icy road and my car began to fishtail. I was belted in and my brother, seated next to me, was not.

By the time we hit the ditch, centripedal force had flung my brother up against the back windshield – along the way dealing me a black eye when his head smashed into my face. I was still seated in the driver’s seat, with enough nominal control of the car to have steered it into the ditch rather than oncoming traffic.

I’ve worn my seatbelt religiously ever since.

greenup April 11, 2008 3:02 PM

I wouldn’t be surprised if it was totally subconcious.

People Accept a certain level of risk, and when they Feel safer, (because they are forced to wear seatbelts) they act in a more risky manner without even realizing it. (like driving faster to make up the time spent buckling them up)

alan April 11, 2008 3:05 PM

Back when I used to drive, I wore seatbelts, not because it made me feel safer, but because they helped keep me awake on long drives. The physical annoyance was helpful in keeping me more alert on the road. (As well as helped to hold me upright.)

Eadwacer April 11, 2008 3:32 PM

I think the link between the safety device and the system has to be more immediate than with seat belts. For example, there was a report sometime last year about cab drivers driving more aggressivly (I know) when their fleet was refitted with ABS braking systems.

Michael Ash April 11, 2008 3:37 PM

When I’ve heard this sort of behavior modification mentioned, it has always been in the context of air bags, not seat belts. It would be interesting to compare the two. I could see why it would happen in the case of air bags but not seat belts. Seat belts are minor and obvious. To the average person, they promise to help you out in a crash but you’re still going to be in rough shape. Air bags are much more magical, and while I don’t want an air bag going off in my face (although preferable to kissing the steering column!) a lot of people may think of them as magical devices which will guarantee a happy outcome.

Seat belts also help prevent accidents by helping the driver keep control of the vehicle, whereas air bags just help mitigate them. This may also change the numbers.

eh April 11, 2008 3:42 PM

Except that now it’s even more dangerous to not wear a seatbelt BECAUSE of airbags. The seatbelt keeps you from getting too close to the bag before it blows up in your face.

tom April 11, 2008 3:42 PM

I started wearing a seatbelt when I bought my 2nd car. The 1st had bucket seats, the 2nd had a bench seat.

I had been driving “Dukes of Hazzard” style on dirt roads. The seatbelt kept me in place as I slid sideways around the corners.

I’m not in high school anymore but I do live in the northeast with snow.

I wear the seatbelt to increase my control of the car so I can prevent the accident that would deploy my airbag. It works.

David Hayes April 11, 2008 3:58 PM

I’ve always thought the Compensating Behavior theory was rubbish. My example is helmet wearing while cycling or snowboarding. While I know the helmet will reduce the risk of mashing my head into a pulp hitting a tree or car at 60mph is still really going to hurt and I know wearing a helmet doesn’t change that. I’d be amazed if people (even subconsciously) changed their behavior to increase these risks, especially when literally risking life and limb

Tangerine Blue April 11, 2008 4:12 PM

This paper investigates the effects
of mandatory seat belt laws on driver
behavior and traffic fatalities.

Actually, it sounds like they studied only fatalities, and are using fatalities as a proxy indicator for driver behavior.

But using seat belts reduces fatalities, so lots of reckless driving won’t result in fatalities, and will be completely off their radar.

I don’t think they can make conclusions about seat belts’ influence on driving behavior unless they study behavior directly or at least include non-fatal accidents.

Jared April 11, 2008 4:38 PM

I think Tangerine Blue has it right. I gave up on the study when I got to the point in the abstract where they declared they only looked at fatalities. It would be easy enough to count all injuries, which is what we ought to be interested in since we want to know about the margins of risky behavior.

I don’t think theory would predict someone with a seat belt starts doing 95 on a narrow mountain road in the rain, but they might take a turn a little quicker than they otherwise would have and end up with a mild concussion.

Clive April 11, 2008 5:11 PM

I’ve been wearing seatbelts as a passenger and then as a driver for almost all my life (I’m a mid-thirties Brit) so that’s my default. I feel extremely vulnerable if I’m in a motorised road vehicle without a seatbelt – even buses and coaches and can scarcely conceive of driving like that at any speed.

This certainly makes me a data point in favour of people taking more risks when wearing a seatbelt – without, I wouldn’t even risk turning the key. (-8

MLR April 11, 2008 5:54 PM

I don’t think it’s a stretch to say that the majority of people use seatbelts regularly. Therefore it becomes an automatic routine. You probably can’t remember what your seatbelt felt like last time you got in your car, because it was something you did automatically, probably while other things were on your mind.

I suppose the reason that seatbelts have so little bearing on the recklessness of drivers is that seatbelts are an established routine. We don’t consciously think about the action, and instead rely on motor memory. We would usually only consider the cost-benefit aspects of the belts if they were drawn to our attention (at our initiation to them or when our routine fails, infrequent for seatbelts).

An interesting side effect of this is that even if we can find this recklessness effect, it’s possible that it would dissipate as users adapt their routine to the system–for simple tasks anyway.

fmackay April 11, 2008 6:13 PM

D. Hayes – your cycle helmet will do nothing to prevent you mashing your head to a pulp – it might mitigate some minor injuries but the rates of death and serious injury are actually slightly higher for helmet wearers. Cause unknown – probably a combination of risk compensation and helmets exacerbating some injuries – the larger effective head size is a) more likely to hit something and b) provides more leverage for rotational injuries (v. bad, broken neck, brain sloshing about etc). But don’t let that keep you off your bike – cycling is a very safe activity – safer than walking, mile for mile, and does not require special protective equipment. You wouldn’t wear a helmet walking to the local store for your groceries, would you?

On seatbelts, as I understand it seatbelt laws do reduce fatalities for car occupants but increase them for pedestrians, cyclists and other more vulnerable road users, so that overall reductions in fatality rates are almost never seen.

moo April 11, 2008 6:17 PM

oh, @Bruce: we clearly need a thread about the security tradeoffs involved in the wearing or not wearing of bicycle helmets. 😛

MPC April 11, 2008 6:54 PM

Looks to me as if the authors did an incomplete job during the experiment. Their dependent variable was fatalities. There are many independent variables that affect that. Cars in general are safer. For example, my Honda Odyssey will literally drop its engine on the road during a crash to prevent it from hitting the passengers. I would love to see an experiment where they measured accident frequency (regardless of accident impact) to determine the factor of risk compensation here. The source data is out there in insurance companies and accident reports, and would be considerably more definitive as to the effects.

Stephan Zielinski April 11, 2008 9:30 PM

With regards to “Jim J”‘s note, specifically “I recall something from my graduate economics course. . .” I have an ominous feeling economics types tend to propagate that one because it fits in so well with what they know of economics. As a concrete example, here’s a paper purporting to provide evidence for the Peltzman effect specifically with regards to car safety:

Okay, they’re extrapolating behavior on the highways from NASCAR accident rates; that’s just garden-variety foolishness. But they also use the difference between first and second place prize money as the primary index of how much winning first place is “worth”. If a physicist picked that number to hang the thesis on, it’d just be sad. From economists– who really should be able to understand that the financial benefits of being a NASCAR driver stem primarily from corporate sponsorships and product endorsements (which are themselves more a function of career than the results of any particular race), making the nominal race purse a gnat’s sneeze worth of the total compensation– it looks to like digging through the raw data looking for a scatterplot with approximately the right shape.

NickB April 12, 2008 12:22 AM

Risk Compensation certainly has some intuitive support: think how much more carefully you would drive if on the centre of the steering wheel a spike were pointing at your chest.

Anon Noob April 12, 2008 4:04 AM

@Anon Techie:
The Compensating Behavior for condom use is best illustrated by examining condom usage in porn. As condoms have become more prevalent in porn, porn has become more extreme.

We see a similar correlation with availability of tissues and Extreme Sneezing.

Erik N April 12, 2008 4:35 AM

I’ve always argued that drivers should be stripped of security devices such as air bags and seat belts. Passengers are innocent, but I don’t have the patience to let drivers learn by error – better let them commit their mistake only once in a life time. If Darwin was right we will evolve into better drivers.

The problem with this point of view of course is that it is not always the driver that is responsible – it may well be somebody else. A device that can determine guilt in the moment of the accident and disable security devices on the guilty should be developed.

SteveL April 12, 2008 7:42 AM

Putting aside whether people change their own behaviour with bike helmets, a Bath University researcher has been collecting data on how others treat you when you have a helmet on. By recording the colour and distance of vehicles passing his bicycle, his data implies that if you have a bike helmet on, vehicles drive closer to you. you get more clearance if you have a long haired wig on.

Anonymous April 12, 2008 1:45 PM

@Erik N
“A device that can determine guilt in the moment of the accident and disable security devices on the guilty should be developed.”

You posted this to the wrong set of comments. You want the “Third Annual Movie-Plot Threat Contest”.

Kjetil Kjernsmo April 12, 2008 2:37 PM

Some claimed that this was the case with avalanche beacons, i.e. radio transmitters that extreme skiers use to speed up search done by peers. Some asserted that it just made people run down steeper faces than they should. This claim has now pretty much been rejected by most practioneers, and most extreme skiers use them nowadays. I certainly do, and I’m pretty sure it didn’t change my behaviour.

Also, I’m a strong advocate of bicycle helmets. I would most likely not have been writing this if it wasn’t because of a bicycle helmet. I’ve done some wild things in my days, but the paradox was that the crash that could have killed me came on a quiet trip on a familiar road with my ex-gf. Again, using a bicycle helmet certainly did not alter my behaviour, but it might be that other’s attitude towards me change.

As an aside, it was quite funny to see that when I came back from having climbed one of Norway’s most challenging peaks, Storen, I was the only one wearing the seat belt on the bus home… I consider it cheap security. I suspect people of just not wanting to think about risks. I’m accepting a lot of risks, but I think I do it consciously.

tobias d. robison April 13, 2008 6:54 AM

This is a terrible study. It does not consider the following scenario, which is nearly untestable: If you make motorcycle riding more safe, some riders will add another equally unsafe aspect to their lives that has nothing to do with motorcycles.

People are not all the same. Some of us DO like a certain amount of stress in our lives, and we might behave in this weird way. Others can enjoy low levels of stress, and will be happy to cycle more safely, period.
– The precision blogger.

GLK April 13, 2008 7:36 AM

I grew up in an era of no car seats, no air bags, and no seat belt laws. The only reason I wear a seatbelt now is because my significantly younger wife’s habits have rubbed off on me. Same reason why I begrudgingly recycle my trash. That, and wearing it is one less thing the cops can fine me for. Back in 1992 I was not wearing a belt while stopped at a traffic light when suddenly I was faced with an incredibly inept driver who would plow nose first right into my driver’s side car door. The impact would shove my entire car into the lane next to me. With only a split second to react I hopped over the console, shut my eyes and prepared for impact. The side of the car was caved in to the point that it bent the driver’s seat pushing it into the console. Consequently I let myself out of the passenger door and walked away without a scratch. Thus proving that belts, like any other safety device, are a compromise. I’m sure if I would have been belted in and suffered a coma or died people would have said, well, he did everything he could.

samsam April 13, 2008 11:01 AM

I like the instructions that came with my kid’s bike helmet: “This helmet cannot protect against all head injuries. Just as an egg can be scrambled without cracking the shell…”
The mental image kept her off the bike for a few days, thus rendering her immune to bicycle-related head injuries.

Dreamer April 13, 2008 11:07 AM

I think that an increased sense of safety has to be establish, not simply the presence of a safety measure. That complicates things a little, but I think it is a little more realistic. In the case of a seat belt or helmet, it seems a bit odd to perceive a sufficient adjustment in the internal risk assessment process to result in an downgrading of precaution from their presence alone.

I do think that compensating behaviour is at least a sound concept, but it seems a little superficial. As others have stated, that simply looking at fatalities isn’t enough to really prove causality for seatbelt use. It is the actual behaviour that matters, so perhaps a study based on driving tests. Building and industrial site accidents would also prove excellent sources of data if the focus was instead on fleshing out Compensating Behaviour as a model, rather than showing any particular safety feature caused more accidents than it was worth.

Stuart Young April 13, 2008 9:38 PM

As mentioned, there are lots of possible statistics that should probably have been compared, such as: Total number of accidents, non-fatal accidents, recorded occurrences of speeding, recorded occurrences of reckless driving, etc.

For such a study, fatalities by themselves are a useless indicator and simply result in the paper becoming a statistic itself; adding to the “number of badly researched papers”.

WarLord April 14, 2008 12:04 AM


Some time ago I read a study of bike helmets and driver behavior.

When the researcher wore his bike helmet cars ran closer to him than when he rode bare headed.

Sort of the same theory, he was “protected” and driver consequently gave less safety space to him and his bike.

Enjoy the journey


bob April 14, 2008 7:17 AM

This study assumes seat belts LAWS == more safety. They are essentially studying the first derivative of seat belt usage. They should have compared actual seat belt USAGE. Not the same thing at all.

Alex April 14, 2008 7:53 AM

Wait while I disable my firewall, AV software, and run as root; then, presumably, I’ll unconsciously optimise my browsing activity for maximum safety! In fact, perhaps I should reboot in Windows for the full effect?

Hmm, then there’s the fuse box; I could bypass all those damn things, and then I’d take so much more care with electricity. (Actually, I’d probably never switch anything on, so that may not support my point.)

Obviously, if I owned a gun, I’d make sure to keep it loaded, cocked, with the safety off, just lying around the place. SCIENCE!

Russell April 14, 2008 9:11 AM

I can believe that – I started riding more safely when I bought a bike helmet. Its presence reminded me that crashing was a possibility. Without the helmet there to remind me, my mind spent more time thinking about other things.

paul April 14, 2008 10:42 AM

Since compensating behavior is (presumably) a psychological effect, it’s only going to last for some fraction of a generation, until people get used to the new default. I’ve worn seatbelts pretty much all my driving life; if I went without one now, I’d probably be so focused on the fact that I was unprotected that my driving safety would suffer seriously.

I haven’t, by the way, seen what seems the more obvious explanation of why motorists give less clearance to cyclists wearing helmets, namely that the helmet is being used as a proxy to signal competence. Someone without it is (a motorist may presume) more likely to be inexperienced or risk-taking, and hence more likely to do something stupid or encounter difficulties that would bring them suddenly into the path of traffic.

Similar signals apply for pedestrians. I will slow down somewhat less and give microscopically less clearance to a roadwork employee in a safety vest than I would to some random pedestrian in the same position next to a car on the roadway. Not because the vest protects them, but because I perceive that they’re less likely to do something stupid and step out in front of me.

Melchior April 14, 2008 12:11 PM

Mandatory seatbelt laws are not about safety, they simply give the cops a reason to pull you over in case you are dumb enough to not wear one.

Orv April 14, 2008 6:26 PM

Seatbelts don’t make me feel safer; they just remind me that driving is a potentially lethal activity. They are the opposite of security theatre, in that they do not actually make you feel safer, just more conscious of risk.

While they nevertheless will substantially reduce risk to your life and limb, in the event of an event that you really don’t want to believe can happen to you.

Pretty good security, I’d say.

jayh April 15, 2008 8:59 AM

“For example, there was a report sometime last year about cab drivers driving more aggressivly (I know) when their fleet was refitted with ABS braking systems”

This is more likely to have that ‘compensating’ effect than seatbelts. For a seatbelt to do something there usually has to be a crash, which people still want to avoid. With ABS, the mythology suggests that you can stop much faster which eliminates that perception of risk.

[Actually in almost all circumstances ABS will NOT help you stop faster. It will occasionally help the panicy driver who mashes the brake pedal]

Kwai Noi April 15, 2008 7:45 PM

I remember some time ago a comedian talking about how airbags should be replaced with a 9 inch steel spike that came out of the steering column on impact if you really wanted people to drive safer.

Just from my own observations of 30 years of driving it would appear that people are driving much faster, and technology has allowed us to distract ourselves with phone calls, GPS, SMS, internet and even TV. So I can see a little bump in statistics from recklessness being lost in all the other causes.

But do we not see more people taking more risk all the time simply in the name of recreation? Look at rock climbing, bungee jumping and base jumping or even the stuff teenagers are doing on skateboards and bicycles these days.

It seems it is only human nature to push the limits of anything we do.

Alex April 16, 2008 7:10 AM

Regarding ABS, aren’t you more likely to have it on a bigger, faster, more expensive and hence more dangerous car?

maxim April 16, 2008 4:28 PM

I think the link between the safety device and the system has to be more immediate than with seat belts. For example, there was a report sometime last year about cab drivers driving more aggressivly (I know) when their fleet was refitted with ABS braking systems.

bigdaddy April 17, 2008 4:49 PM

I don’t like to tell my age but I was a police officer in 1955 and the cars all had seat belts, I still never get behind the wheel without the harness.
Has saved me several times just in manuvers that would cause you to leave the position behind the wheel!!!

Jeffrey Austin White June 25, 2008 8:59 PM

I wonder how many deaths are attributed to users who can’t seem to figure out how to use the seatbelt. Apparently that is why instructions are still provided prior to airline flights.

Jeffrey Austin White June 25, 2008 9:01 PM

I wonder how many deaths are attributed to users who can’t seem to figure out how to use the seatbelt. Apparently that is why instructions are still provided prior to airline flights.

Alaban August 1, 2008 12:43 AM

Seat belts are minor and obvious. To the average person, they promise to help you out in a crash but you’re still going to be in rough shape.

oldbat March 28, 2014 10:56 PM

I read so many variations on these pseudo-libertarian-themed safety studies, all purporting to reveal some paradoxical truth, but unfortunately most link back to a tiny number of outdated or discredited studies which often didn’t even relate to the subject people now are ascribing to them.
What this means is for the most part, it’s horse dung. Folklore, old wive’s tales, mythological or apochryphal tales used to insinuate a certain lesson, moral, political slant, etc.
These pseudo-psychological reports have not been widely accepted, but then neither are the reports of lizard people and the efficacy of orgone generators.
The most basic broad accident statistics demonstrate clearly the reduction by half, highway fatalities in the years since seatbelts and vehicle safety designs have been mandated. No compensation needed, here.
It is striking even to the casual viewer, whenever an accident hits the local news, how many victims are described as “was not wearing a seat belt”,”victim was thrown from the vehicle”, versus “passenger who was wearing seat belt was treated and released from the hospital”.
Virulent, “freedom-loving” propaganda is insane, when it must stoop to inventing lies about such fundamental issues such as traffic safety.

Leave a comment


Allowed HTML <a href="URL"> • <em> <cite> <i> • <strong> <b> • <sub> <sup> • <ul> <ol> <li> • <blockquote> <pre> Markdown Extra syntax via

Sidebar photo of Bruce Schneier by Joe MacInnis.