Smeed’s Law and motorcycle fatalities

We’ve looked at the various pieces of the motorcycle safety puzzle and found that they all—without exception—have failed to bring the death toll down but as more riders practice them the death and injury toll goes up.

It’s time, then to explore other things that might affect the crash rate of motorcycles in America. Some of these readers have referred to—and we’ll look at them more closely. Some of them may seem quite far-fetched and some might be rather offensive. Yet, since the usual answers haven’t solved the puzzle, it’s appropriate to explore other factors—no matter how unpalatable—in case they may in part or in concert led to safer roads for riders.

We start with R.J. Smeed’s “Law” which was first published in 1949. It states that as the number of automobiles in a country increase so do fatalities in a predictable way: the number of deaths equals .0003 times the two-thirds power of the number of people times the one-third power of the number of cars.[i] After that point, road fatalities begin to fall off and then level off at a much lower point.

Despite safer cars, Smeed’s Law is still basically true in all developing countries. For example, it held true in the USA until about 1966—and his formula for the decline of traffic fatalities is very close to what has actually happened.

His friend, the eminent physicist Freeman John Dyson, wrote, “It is remarkable that the number of deaths does not depend strongly on the size of the country, the quality of the roads, the rules and regulations governing traffic, or the safety equipment installed in cars. Smeed interpreted his law as a law of human nature. The number of deaths is determined mainly by psychological factors that are independent of material circumstances. People will drive recklessly until the number of deaths reaches the maximum they can tolerate. When the number exceeds that limit, they drive more carefully. Smeed’s Law merely defines the number of deaths that we find psychologically tolerable.”[ii]

Of course, in 1965, Ralph Nader’s book, Unsafe At Any Speed, was published which both captured the general public’s growing frustration with traffic fatalities and exacerbated that frustration. From the mid-Sixties on there was a massive push for safer design, safer roads and safer crashing. Iow, Smeed was right about the linkage but assumed it would take more cars and deaths to get to the point we could no longer psychologically tolerate the death toll.

It’s true that motorcycles can’t be made as objectively safe (crush zones, front and side air bags, etc.) as cars—but then that’s true for bicyclists and pedestrians as well and their death rates have dropped in the past ten years while motorcyclist fatalities rose—and rose and rose outpacing registrations.

When it comes to automobiles and perhaps bicycles[iii], there’s not just a correlation but some kind of subconscious process at work that first allows the death toll to rise and then, eventually, lowers it.

But the key here is that drivers keep driving—they just drive safer.

The question is: does Smeed’s Law work for motorcycle registrations and rider deaths?  I’ll leave it to anyone who’s better at math than I to do the math but I do wonder: How can we as riders still “psychologically tolerate” the soaring death toll?

But here’s this—even if it does, it’s a little different when it comes to motorcycles:   The past 11 years is not the first surge in motorcycle registrations and fatalities in the USA. The most recent registration surge ended in the early 1980s and fatalities topped out in 1981. The death toll began dropping and bottomed out in 1997—even though registrations had begun to increase a few years earlier.

While 29 states either dropped or adjusted universal helmet laws during the 1970s while fatalities were rising, the laws weren’t reinstated yet fatalities dropped. From 1973-2001, 1.6 million were trained and all states began to require motorcycle licensing—and most were trained as fatalities were falling.

But the death toll did drop beginning in 1982—and so did registrations and then registrations started to go up in the early 1990s—and fatalities followed suit in 1998.

However since 2002, the Motorcycle Safety Foundation claims over 2 million have been trained—and yet fatalities have exceeded the height of the late 1970s-1981 surge in rider deaths.

Today, EMS response time is better than it ever has been, medical procedures are more effective and traffic system design has concentrated on safer roads and intersections. While this has brought about reductions in auto, bicycle and pedestrian deaths, some of that loss was simply transferred over to motorcyclist deaths.

Iow, just as with automobiles, Dyson’s words could be applied to motorcycles. It appears “the number of deaths does not depend strongly on the size of the country, the quality of the roads, the rules and regulations governing traffic, or the safety equipment.”

In this way, Smeed’s Law might be true but in a different way than with cars. When it comes to autos, people are sickened by the death rate and demand change as a nation of drivers—but they keep on driving and registrations keep on going up.

But motorcycling doesn’t behave the same way: in the past three cycles, registrations peaked before fatalities did—but unlike Smeed’s Law predicted, registrations did fall off.

Iow, while drivers either behave more safely or there are changes to design, roads or safety measures are brought to bear, this doesn’t happen with riders—yet the fatality rate still drops. But so does registrations.

It could be that individual riders no longer believe that riding is safe for them and give up motorcycling—and thus increased motorcycle “safety” is really attrition. Which doesn’t make motorcycling safer at all.

[i] Smeed, R. J. Some Statistical Aspects of Road Safety Research. Journal of the Royal Statistical Society. Series A (General), Vol. 112, No. 1 (1949), pp. 1-34.

[ii] Dyson, Freeman. “Part II: A Failure of Intelligence” Technology Review

[iii] Hakamies-Blomqvist, Liisa and Mats Wiklund, Per Henriksson. Predicting older drivers’ accident involvement – Smeed’s law revisited. Accident Analysis and Prevention 37 (2005) 675–680.

Explore posts in the same categories: History, Legislation, Motorcycle crashes, Motorcycle fatalities, Motorcycle helmet use, Motorcycle injuries, Motorcycle legislation, Motorcycle Safety, Motorcycle Safety Foundation, Motorcycle Training, Uncategorized

20 Comments on “Smeed’s Law and motorcycle fatalities”

  1. gymnast Says:

    As I speculated earlier, the “tip point” seems to coincide when an annual death rate on the order of a ratio of approximate 1 death per each 1000 registrations is reached.

    Your essay seems to be along the the same lines with additional additional concepts that tend to lead me to believe that you are “on to something”. While national motorcycle registration and crash fatality figures prior to 1975 tend to be somewhat lacking, figures for individual states may be available that lend validity to the construct. Even though registration figures may be lacking for the early years of motorcycling, I have a feeling that annual sales figures of motorcycles in the us dating back to the 1920s are a part of the MIC’s archives. Reliable annual fatality figures may be a bit more difficult if not impossible to obtain for the “early years”.

  2. wmoon Says:

    I was thinking of that when you had first raised the ratio and I knew I was working on the Smeed Law thing. What I find fascinating is that there may be “social” laws on the order of natural laws (natural social laws) that we’re not yet really aware of–but if there are, then real change has to take those into account–but they may also dictate what real change is possible.

    I wish that MIC/MSF could see how many of the things I raise would better their interests if they cooperated…sigh…

    What I find most important in the Smeed entry is the difference in how drivers and riders handle that point of psychological revulsion–however unwittingly: drivers keep driving and change in some ways both external and internal while riders simply quit riding–and thus repeating the cycle when motorcycling becomes popular again.

  3. Dave B Says:

    “I wish that MIC/MSF could see how many of the things I raise would better their interests if they cooperated…sigh…”

    That would require intelligent life forms being employed at the MSF. As we all know, the Prime Directive over there is helping to sell motorcycles. Getting involved with anything that may interfere, deter or hurt that PD is a no-no.

  4. George Says:

    Great post – I tweeted you from utahbikelaw

  5. DataDan Says:

    I recall Smeed’s Law from Tom Vanderbilt’s recent book, Traffic. While I don’t dispute the observations, Dyson’s interpretation seems to have a paranormal “hundredth monkey” quality to it. It can also be interpreted as adjustment to risk, ala risk compensation theory.

    We respond to perceived ambient risk by making adjustments to bring it down to an acceptable level. On a very short time scale, we may slow and increase space cushion when approaching an oncoming vehicle waiting to turn left. On a greater time scale, we make adjustments to our riding based on our overall perception of the risk of motorcycling. And those perceptions are influenced by crashes we hear about from friends and on the news.

    A good example of this can be seen in the percentage of unhelmeted crash injuries in Florida following helmet law repeal (see chart linked in my response to your previous post). Unhelmeted injuries spiked to 54% of all injuries immediately after repeal but fell to 36% over the next 5 years. Either helmet use increased or those riding unhelmeted got more cautious–or both. Riders responded to evidence of their vulnerability in a rational way: by reducing risk, which is seen in the declining injury rate.

    Though I see the long-term ups and downs of motorcycle popularity as driven by population demographics, you make an excellent point about selective attrition. Hurt and MAIDS both found strong correlation between inexperience and crashing. But more experienced riders, who crash significantly less often, are actually self-selected for some sort of innate motorcycling abilities (whatever those may be). In their “freshman class cohort”, those who started riding at the same time, many gave it up after discovering–one way or another–that they weren’t cut out for it. So to some extent (though certainly not totally), riding is safer for more experienced riders because they’ve survived a selection process that eliminated less safe riders.

  6. wmoon Says:

    DataDan, I’m in the midst of Traffic and really enjoying it. Yes, I told all of you that there may be things some find objectionable–and Smeed’s Law if you think about it–as you have–is one of them: we don’t want to think there’s some kind of “hundredth monkey” sort of thing. But if we’re not willing to look at the possibility that there may be forces that we do not understand, that go beyond simple explanations (or have such simple ones that we simply don’t comprehend them) then we’re not doing our job of exploring other options when the ones that should work aren’t. You raise a sort of Darwinian thing–and that’s something we need to look at more closely as well.

    I will be getting to risk compensation soon–so we’re thinking along the same lines, I’m thinking. If you look at national stats back to ’75 you see the same thing as in FL and LA. The mid-70s till 1980 were a time of many universal helmet law repeals and fatalities shot up (though I haven’t looked for helmet v. unhelmeted data). Yet, after a few years then fatalities began to fall. Is that because risk compensation (or homeostasis or off-setting behavior) adjusted to a new perceived level of risk? I don’t know. I do know we can’t assume what the answer is, though and be objective or responsible without more evidence and exploration.

    I find it interesting that you went where my mind went–we have to factor in the generational demographics, factor in the recession of the early 80s and the others, etc. Which is why I’ve brought up all those things in past months. What I’m trying to do is build a more comprehensive picture that factors in all those kind of things. I appreciate your contributions–keep them coming.

  7. wmoon Says:

    But that’s the thing, Dave: if they didn’t have their heads up their short-term quarterly pennies per share collective asses they’d see that if they took a long-term approach with yearly growth expanding more slowly but with long-term riders as a result, that their companies would sell far more motorcycles.

    They are colossal idiots that they cannot see that of all forms of transportation only two–bicycling and motorcycling depend to a huge degree on a public perception of acceptable risk and particularly a human acting as both attractor and validator to these worlds. If people both saw riders and bicyclists in greater numbers–and the general perception was they were safe and kept safe–then more would ride and thus more vehicles would be sold. But when the general public gets the overall impression through a variety of sources–but most importantly personal knowledge of someone they know (or they know someone who knows someone) that helmeted, trained riders aren’t safe… well, you see the problem.

  8. Jeff Brenton Says:

    I can’t help thinking that maybe, just maybe, there is a finite number of safe, skilled riders/drivers in the general population, and that, unfortunately, any increase in the number of riders brings a higher percentage of “less safe” and “less skilled” riders into the fold that safe, skilled riders.

    Let’s say there are 100,000 riders, of which 30,000 would into that “safe, skilled” group. If the total population increases by, say, 10% to 110,000, that doesn’t necessarily mean that there are now 33,000 “safe, skilled” riders… it could mean there are only 30,500, because the pool of new riders might not have the same mix as the original group.

    A factor like that could explain both sharp increases and sharp decreases – as new riders flood in, a lot of them crash and die. When the sales swell abates, it’s the less-skilled riders who drop out first, and the crash rates fall.

    Supporting anecdotal evidence is the abundance of bikes that hit the market with low mileage and crash damage. The rider scared (or injured) themselves enough to drop out, after being a part of the increasing registrations.

    Just a thought…

  9. wmoon Says:

    Jeff, It’s not that I’m opposed to your theory–I’m just wondering what you’re basing the idea that there’s only X number (whatever that would turn out to be) safe/skilled vehicle operators and that percentage wouldn’t vary. That might apply to drivers since almost 100 percent do drive–but motorcycling requires not only a specific choice to ride but a decision to accept a higher level of risk (even if woefully low compared to the real risks). And is this specific to just motorcyclists or other groups. If so, are X number of bicyclists, truck drivers and pedestrians poorly skilled?

    The anecdotal evidence you suggest can be evidence of many other things than poor skill–including a re-evaluation of the real risks and their estimation that they can’t control it–but not necessarily because they feel their skills are deficient. So if it is–then do the number of used cars for sale or bicycles mean people can’t drive ride a bicycle and give it up?

    So if you’re going to put this out there–what would be a more solid basis for arguing it?

  10. Jeff Brenton Says:

    It is hard to say how you could prove or disprove it. The idea is that the largest percentage of those who are best equipped (mentally) to deal with the risks are already in motorcycling, and have been, long-term. Those that move in and out of it on a cyclic basis are NOT as well equipped, so they show up more frequently in crash statistics.

    These “cycle riders”, if you will, have a double disadvantage of “maintaining low experience”; They never become proficient, despite (sometimes) years of riding. The Hurt report found a correlation between low time on the current bike and likelihood of crashing; could this also be a problem for someone who has owned their bike for 10 years, but only put 5,000 miles on it in that time?

    In aviation, there is the case of pilots who have “800 hours” experience, yet make errors you would expect of a student pilot. I’ve referred to it as being “50 hours experience, 16 times”. They never reach a real level of proficiency, and are over-represented in the crash statistics. This is why the NTSB classifies recent pilot experience in the crash data (last 30/90/180 days), in addition to total experience.

    I know riders with dozens of years (and hundreds of thousands of miles) of crash-free riding, whose motorcycle skills would be found wanting, if tested, yet they stay out of trouble. I know riders with excellent motorcycle skills that can’t go to the market without having at least one close call. I attribute this to judgment, or lack thereof.

    Those with poor judgment make poor riders. And yet, their poor judgment might make them susceptible to becoming riders, without considering the risks. They also make poor drivers, but they have 40 years of safety equipment to protect them from that poor judgment in a car.

    Despite our rate of crashing and killing ourselves, bike crashes are still “under the radar” of most people. They hear about a dozen car crashes a day, but maybe one or two bike crashes a week. It could be that the perceived danger doesn’t outweigh the image people associate with riding.

    Those that look upon it more as a form of transportation than a “lifestyle statement” are, quite possibly, better equipped, mentally, to deal with reality. Those riders with hundreds of thousands of miles of crash-free riding almost always consider it to be a “fun way to get there”, but it’s still just a “way to get there”, rather than a statement about “who they are”. And maybe that’s the difference.

  11. DataDan Says:

    Here’s a variation on Jeff’s speculation about a limited number of potentially “safe, skilled” riders in the population that leads to a similar result.

    An interesting paradox about motorcycling is that it *attracts* risk takers, but it *rewards* the more risk-averse. Those who take the biggest risks are weeded out early, while those who take less risk gradually develop skills and confidence and may go on to enjoy many years on a motorcycle. So, in my version, it isn’t a limited number of safe riders in the population but a small percentage of those who start out. In a given freshman class of riders, many will be out after just a few years.

    When motorcycling is growing fast, there will be a high percentage of newbie riders in the population. And, per my speculation above, there will also be a relatively high percentage of high-risk ex-riders-to-be. But when the sport is in decline, the newbie percentage will be low and there will also be fewer high-risk, short-career novices.

    That would help explain both the falling fatality rate per registered bike from the mid 1980s to mid 1990s–when registrations dropped by one-third–and the rising fatality rate from 1997 to 2005–when registrations nearly doubled.

  12. wmoon Says:

    Jeff, I don’t even know where to start other than to say you’re making a lot of unsupported assumptions and are using a finding from the Hurt Report (that was also found in MAIDS) out of context in a way never intended by Hurt, at least (as I can speak to what he told me while I haven’t talked to those were more deeply involved in MAIDS).

    While it’s true, as Steve Garets has said for years, that a rider can have 10 years of experience or 1 year of experience 10 times, even so, you’re taking that truism in a way that cannot be supported by Garets or you or anyone–there simply is no evidence other than your own evaluation. So it may or may not be true. However, studies have suggested that riders are no more “risk-takers” than drivers nor that their judgment is lacking–and it has found their hazard-awareness is superior. Nor, despite what you believe, has experience been found to be a slam-dunk for safe riding. Nor has exposure (miles traveled) been found to be a true-test. In fact, exposure is curvilinear.

    Nor do I think the lifestyle v. transportation designation is anything but an unfair and erroneous generalization that not only isn’t supported by statistics but smacks of prejudice and bias.

    Once again, Jeff, I urge you to dig in and find some evidence or at least consider more things in what appears to be a very simplistic model. It may be that you’re right–but you have to have more than your very limited experience to back it up to carry enough weight.

  13. wmoon Says:

    DataDan, Actually studies have found that riders are no more risk-takers than drivers. And there’s simply not enough evidence to support rewarding the risk-adverse–in fact, as we’ve been exhaustively looking at, the risk-adverse (those that take training, get licensed, wear gear and ride sober) aren’t rewarded–in fact, they’re dying in greater numbers than ever before. A huge number of riders (including me) want to believe what you and Jeff say–and most would share your views–but the fact is it isn’t working.

    Nor is there any proof that “those that take the biggest risks are weeded out early.” In fact, the evidence seems to point the other way–for ex. the California studies. Nor does your model take into account the lifecycle of riders who give up not because of danger or crashing, etc. but because of other life demands–and themajority of the Boomers career/family years hit during the decline from 1982-1998.

    I don’t think it’s fair to categorize returning riders as “newbies” either nor is there any evidence they are either less-skilled nor more accident-involved. In fact, some studies have found the opposite.

    I find that both you and Jeff are making assumptions that, I humbly suggest, seem to confirm your own experiences and beliefs and may be affirmation biases. Which is not to say you’re not correct–but that you and Jeff are making unsupported assumptions that need something (and it doesn’t necessarily have to come from motorcycling to be true) rather than really looking at the situation we face.

  14. DataDan Says:

    My assertion is indeed speculation, but I stand by it.

    When I say that “motorcycling attracts risk-takers” I mean that those who take up riding have a higher tolerance for risk than the average Jo. Not that they are daredevils; just that they are more comfortable than most folks would be in an activity that is regarded as dangerous by the public. Nor do I mean that on a trip from Point A to Point B the typical motorcyclist will engage in behavior that is more likely to result in a crash than does the average driver. As an avid student of motorcycling risk, I would be very interested in references that contradict my speculation.

    However, I am not asserting that motorcyclists are uniform in their risk-taking preference. There is still quite a range among those who take up motorcycling. But the ones who are still riding after 5-10-20 years either were at the more cautious end of the spectrum all along, or learned how to limit their risk by expressing their wild side on the track or in the dirt (both safer environments than the street), or just by exercising self-control.

    I certainly don’t have proof of my assertion that motorcycling “rewards the more risk averse”. That’s a personal observation about many riders I’ve met over the past few decades. But I can offer at least one factual snippet that suggests a difference between those with higher and lower tolerance for risk.

    I would argue that young riders on Suzuki GSX-Rs (their 600-750-1000cc high-performance sport models) tend to be high risk-takers. OTOH, young riders on Suzuki SV650s (generally regarded as a good starter bike) tend to be lower risk-takers. That’s an observation from meeting riders both on the net and in real life. I read more bravado from Gixxer riders, and I get into more discussions of safety issues with SV riders.

    In 2008, 560 Californians died on motorcycles. Of those, 60 were on GSX-Rs, 46 of them under age 30. But only 6 died on SV650s, only 2 under 30. That’s a huge difference: 46 under-30 Gixxer deaths vs. 2 under-30 SV deaths.

    Of course, that’s not much more than anecdotal. But it is a counterpoint to your claim that the risk-averse are dying in greater numbers.

  15. Jeff Brenton Says:

    The fact that there is no way to prove/disprove my idea is why I really don’t like it… And, should it be true, there’s not a lot you could do about it, which galls me as an educator. Short of severely restricting who is allowed to ride, only after they’ve proven themselves to be excellent drivers. But it IS a possible explanation.

    In my version of this, it doesn’t come down to whether someone is a “risk taker” or “risk-adverse” – they simply do not understand the risks. The driver who has 100 near misses a year doesn’t usually think, “Hey, it’s my driving that sucks,” but is more likely they think, “If I didn’t drive as well as I do, all these idiots would have killed me!” They don’t see their behavior as contributing to the risk, so there is nothing for them to fix.

    Those studies did comparisons of relative risk-taking and awareness between drivers and riders, or were they dealing with “absolutes”? Relatively speaking, there might not be a difference between the two groups. However, the consequences of bad judgment are not the same between them. A judgment made to take a gap that isn’t big enough could be considered the same “risk”, but might result in a dead or badly injured motorcyclist vs. an inconvenienced driver.

  16. wmoon Says:

    We’re going to be talking soon about some of those studies and why we drive/ride the way we do.

  17. Steve Jones Says:

    I think the idea that, on average, motorcyclists assess personal risk differently to those in cars is almost certainly the most important factor involved here. In developed countries, there are relatively few compelled to travel by motorcycle, rather than car, on purely cost grounds. That tends to mean that there are far more leisure motorcyclists or people that choose that method of transport for speed through traffic, sensation and other reasons beyond mere utility of transport.
    I simply think that in many such cases the perception of risk and the psychological rewards are such that it naturally attracts disproportionate numbers of risk takers. Almost by definition, if the safety was a high priority, nobody would choose to ride a motorcyle on a public road. They are inherently far more vulnerable than are cars (crash protection, sensitivity to poor road surfaces, visibility to other road users – this list goes on).
    Smeed’s interpretation of his own law was that the public would somehow have an acceptable level of risk and, maybe subconsciously, drive to that limit. Whether he believed that the perception and acceptance of risk could be affected through social measures, I’m not sure. However, I’m fairly certain that this is the case – mindsets have been changed enormously in other areas; smoking, child safety, pollution, and I don’t think there’s any reason why this isn’t the case for driving. Simply look at the public acceptability of drink driving for one. It also goes a long way to explaining why there are radically different fatality rates in different countries despite similar levels of technical development.
    With motorcyclists, I suspect strongly that it retains a disproportionately high number of natural risk takers of which I see evidence every day on by commute into work.

  18. wmoon Says:

    Steve, If I were you, I wouldn’t assume that those who do not put as high a value on safety as others are inherent risk takers. In fact, risk compensation studies show that people who use safety measures take on more risk naturally–simply because they believe they AREN’T taking on more risk. And those who eschew safety measures don’t seek risk–rather they feel they don’t need the safety measures because they are behaving within acceptable risk.

    Iow, people who ride believe that they have some other way of mitigating the risk that off-sets the motorcycle’s greater vulnerability. This has been found true in sky-diving, boating, parent’s supervising kid’s on obstacle courses, etc.
    Iow, those who look to riders at greater risk-takers are missing a fundamental truth about human nature–and doing so, they can exacerbate the problem. Iow, making riding “safer” by building myths about helmets, gear and training makes people who would never take on that amount of risk believe it’s safe to do so.

  19. Steve Jones Says:

    If you read what I said again you will see I referred to perception of risk was different. Just because you don’t perceive yourself as a risk taker, doen’t mean you aren’t one. That’s very different to thrill seekers (which is what I think you mean). Those people do, deliberately, seek out a degree of risk for the high. When I used to ride motorcycles, then I loved the thrill of it too, but as a large number of motorcyclists find out, the “it won’t happen to me” belief typical of the young (and some not so young) turns out to be untrue.

    As for not wearing crash helmets – well, anybody who does so on the public road is taking a considerably increased risk. You may be the safest rider in the world, but you can’t control the actions of others.

  20. wmoon Says:

    Steve, What you seem to fail to appreciate is that the perception of risk–and the judgment that someone takes undue risk is inherently subjective. There is no objective standard for anything, any activity nor are our estimations of probabilities objectively based as they, too, are subject to our assumptions/beliefs, etc. The “it won’t happen to me” is inherent to the human psyche–even for the little old lady from Pasadena.

    And who said anything about not wearing helmets? I sure didn’t. As I’ve made clear over and over, I choose to do so. My futile efforts have been to get riders and others to recognize that they are not effective in the majority of cases from preventing injury or death and that the sole answer is to stay out of situations where we have to find out if, in our crash, they will be effective.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: