home        facilities


This is a just-completed portion of my new book Bicycle Transportation: A Handbook for Cycling Transportation Engineers, that is now in press at M.I.T. I publish this now because the FHWA's Manual has just been issued.

Copyright John Forester, 1994


In 1994, having been defeated in its attempt to prove that bike paths made cycling safe, and having failed in its attempt to prove that bike lanes made cycling safe, the Federal Highway Administration issued a manual "Selecting Roadway Design Treatments To Accommodate Bicycles." This manual specifies the width of the outside lane and the use of a bike-lane stripe for different types of road and cyclist. This document is significant in different ways. If implemented, it will work against cyclists by both installing more dangerous facilities and by reducing the incentive to learn how to ride properly and safely. It presumes that most cyclists will always be incompetent and ill-informed.

It is significant in another way also. In it, government gives up all attempts to demonstrate that bikeways make cycling safer or easier. This is a tacit recognition that these attempts, stretching over twenty years, have failed utterly. While this document quotes a statement about the safety effect of bike lanes, that statement is twenty years old and even when new had no scientific basis. In fact, this document, both in the quotation just mentioned and elsewhere, tacitly admits that bike lanes make driving bicycles and motor vehicles more difficult and more difficult to learn. Therefore, the practices that this document specifies have no justification whatever. Yet government still persists in its policy against cyclists.


The ostensible reason for the Manual is to develop cycling transportation by providing a bike-lane system that persuades beginners and parents of children that the system makes cycling safe for people with poor cycling skills. Since the attempts to develop scientific support for all substantive parts of that chain of reasoning (that is, excepting the psychology of fear) have failed, that reasoning is highly suspect. Therefore, the most likely reason for FHWA's issuance of this document is that it continues the FHWA's traditional policy of trying to keep cyclists out of motorists' way.


The FHWA justifies its policy by classifying cyclists into three groups. What it calls advanced cyclists are those 5% who know how to ride in traffic. Basic cyclists are "casual or new" riders who don't know how to ride in traffic but trust that bike lanes enable them to do so safely. Children are those under thirteen who don't know how to ride in traffic and whose parents believe that bike lanes make cycling safe for them. Since the FHWA believes that B and C cyclists are very similar, the FHWA groups them together for traffic purposes. Thus there is the A group that, by the FHWA's claim, constitutes 5%of present cyclists and the B/C group that constitutes 95% of present cyclists. The rest of the policy then assumes that the B/C group will continue to be the large majority for whom the entire system must be designed. In effect, the FHWA advocates dumbing down the cycling traffic system to suit the desires of the least competent possible users. That policy nicely serves, as it has for decades, to promote the highway establishment's major cycling interest, its desire to prevent cyclists from delaying motorists.

The FHWA conceals this policy with a smokescreen of words that mendaciously appeal to competent cyclists. "Group A bicyclists will be best served by designing all roadways .... [with] wide outside lanes on collector and arterial streets ... [or] usable shoulders on highways." That would be fine if this were done, but the planning procedure puts this so far down on the priority list that the bike-lane system will be completed first, even on these streets, if the planners think that any B/C cyclists want to use them. The manual says, "The recommended design treatments for group B/C bicyclists should be considered the DESIRABLE design for any route on which this type of bicyclist is likely to ride."


The Manual expects the bicycle transportation plan that other laws require. As always, the assumption is that cyclists, except for that 5% of Group A cyclists, should be traveling on the bikeway system that government has designated. Commendably, saying that cyclists want to go to the same places that everyone else wants to go to, the routes for this plan are to serve the same areas that motorists travel between. The amount of bicycle transportation along a corridor should be estimated as a proportion of the motor traffic along that corridor. However, that bicycle traffic is not to be served with the fast, direct route that motorists choose to use. Instead, a route that appeals to beginners and parents of children must be chosen, if at all possible, and all streets along the chosen route that are not quiet residential streets must be bike-laned. The route shall be designated as a bicycle facility.

All points to which people would want to go should be accessible from the bikeway system. Accessibility is defined as the distance between the point and the bikeway. The following sentence, "No residential area or high-priority destination should be denied reasonable access by bicycle," implies that reasonable access is only by bikeway.

After the facilities for the 95% of the cycling population who are beginners and children are completed, if that ever occurs, something may be done for the remaining 5% of cyclists who are considered competent to ride in traffic. Since these are a small minority of a minority, and since the government has already provided a cycling system, and since the remaining streets that these cyclists find useful are also those which many motorists find useful, the probability of getting the outside lanes widened or even getting the traffic signals reworked is minuscule, non-existent.


The specific cross-sectional design to be used is selected from a table whose categories are: Rider Type (A or B/C), Roadway Type(Urban or Rural), Average Daily Traffic Volume (< 2,000, 2,000-10,000, > 10,000), Average Motor Vehicle Speed (< 30, 30-40, 41-50, > 50 mph), Large Vehicles (< 30/hour, > 30/hour), Adequate or Inadequate Sight Distance. The design features specified for Group A riders are wide lanes (urban) or shoulders (rural) on all but quiet residential streets. The design features for Group B/C riders are bike lanes (urban) or shoulders (rural) on all but quiet residential streets.


How does the FHWA justify its recommendations? Supposedly, there is a research paper behind the manual but, even after repeated requests made to the contracting officer in the FHWA, I have not been sent it because it has not been released. The primary author, Bill Wilkinson, announced the results of his study in his newsletter, ProBike News, long before his contracting officer had even received his research plan. The whole thing smells fishy, as if the research paper were concocted to suit the desired results.

The manual specifically disclaims any scientific support for its design recommendations. "Determining these ranges was difficult; there is little in the state of the practice to go by, and there is tremendous variation in prevailing conditions." In short, this manual has been created by copying the guesswork of other people.

The manual footnotes the research paper as providing support for three statements.

[1] "There is some evidence to suggest that the disruption in traffic operations associated with bike lanes is temporary. Over time, both bicyclists and motorists adapt to the new traffic patterns, learning to look for each other and effect merges prior to intersections."

[2] "Wide curb lanes have three widely accepted advantages. They can: [a], accommodate shared bicycle/motor-vehicle use without reducing the roadway capacity for motor vehicle traffic; [b], minimize both the real and perceived operating conflicts between bicycles and motor vehicles; [c], increase the roadway capacity by the number of bicyclists capable of being accommodated."

[3] "Field studies carried out as part of the research for this manual indicate that bike lanes have a strong channelizing effect on motor vehicles and bicycles. The CALTRANS Highway Design Manual describes this effect very clearly.

'Bike lane stripes are intended to promote the orderly flow of traffic, by establishing specific lines of demarcation between areas reserved for bicycles and lanes to be occupied by motor vehicles. This effect is supported by bike lane signs and pavement markings. Bike lane stripes can increase bicyclists' confidence that motorists will not stray into their path of travel if they remain in the bike lane. Likewise, with more certainty as to where bicyclists will be, passing motorists are less apt to swerve towards opposing traffic in making certain they will not hit bicyclists.'"

The 24 other footnotes don't cover any scientific evaluation of the engineering or safety effect of bike lanes; only the political effects are given a pseudo-scientific coverage. Since all previous attempts at scientific justification have failed, the three statements above are then the sum total of the FHWA's 20-year attempt at scientific justification for its bike-lane policy. It makes most sense to consider the last of these first.

The quoted statement that bike lanes have a strong channelizing effect was written almost twenty years ago and was put into the California bikeway standards to suit the motoring establishment over the objections of cyclists (which were presented by me). While this paragraph is intended to justify bike lanes, and is commonly taken to do that, as Wilkinson believes it does, its actual meaning explains exactly why bike lanes complicate the driving task and produce dangerous driving errors. At the time of writing the effect was deduced and described, but had not been directly measured in a scientific manner.

The first field study of this effect was made by me about fifteen years ago and was presented by me to the Bicycling Committee of the Transportation Research Board. As you will read in the next section of this chapter, this paper was rejected precisely because I demonstrated that the strong channelizing effect caused both cyclists and motorists to commit many more driving errors at intersections. In bikelane areas, cyclists turned left from the curb lane, generally without looking behind, motorists turned right from outside the bikelane, and cyclists got on the right-hand side of motorists who could, and frequently did, turn right. In addition, where the bike lanes were impeded by many stop signs, cyclists ran stop signs without either slowing or looking. Ken Cross had already shown, in the absence of bike lanes, that these errors cause 30% of car-bike collisions. (See the chapter on Accidents for details of his study.) I stated that then it was reasonable to conclude that bike lanes caused car-bike collisions, which was the statement that the Bicycling Committee refused to permit in a paper to be given in its meeting.

The old California argument that the FHWA has adopted clearly argues that bike-lane stripes are good because they keep motorists and cyclists in their "own" areas. As the analysis in this book shows (and this knowledge has been known for twenty years, was known when that argument was written), cyclists and motorists cannot have different spaces. It is physically impossible without separation by either height (overpasses) or by time (separate traffic signal phases). Attempts to define separate spaces, such as by bike-lane stripes, simply produce conflicts between cyclists and motorists that produce dangerous driving errors by both types of driver.

The last sentence in California's paragraph is intended to imply that the installation of bike lanes prevents motorist-motorist head-on collisions caused by motorists swerving across the center line because of the presence of a cyclist on the road. This sentence was originally written in stronger form. I remember the representatives of the California Highway Patrol talking up this threat time after time, but it was merely a big bogeyman created to frighten unsophisticated motorists. Nobody ever documented such a collision. Since such collisions did not exist, I pointed out that to say that bike lanes reduced a type of collision that did not exist was lying. Therefore, the sentence was reworded to merely stir up the fears of motorists about the dangers that cyclists caused them by using the roads, without actually committing a demonstrable lie.

The statement that drivers learn to adjust to the new traffic pattern required by bike lanes is a specific admission of the truth of my argument. All that Wilkinson says he has now discovered is that motorists and cyclists learn, over time, to compensate for the added difficulty and danger. That is no justification for creating the difficulty in the first place. Since lack of driving skill is the primary substantive justification for the FHWA's policy, it is plainly absurd to defend its policy on the grounds that users will, in time, learn to handle the more difficult driving task that it creates.

The statement on the advantages of wide curb lanes contains no recent discovery (I was writing the same things fifteen years ago.) and these advantages demonstrate that wide curb lanes provide all the advantages of bike lanes without the disadvantages.

The use of motor traffic volume on a road to specify the presence and width of a wide curb lane or a bike lane is clearly incorrect. If the volume of motor traffic matters, it is the volume in the outside lane, adjacent to the cyclists. The volume in lanes further left is irrelevant. In justifying consideration of the volume of motor traffic, the manual says, "Higher motor vehicle volumes represent greater potential risk for bicyclists and the more frequent overtaking situations are less comfortable for group B/C bicyclists unless special design treatments are provided." The words "greater potential risk" are weasel words that carry fear without meaning, thus increasing the fears of the B/C cyclists.

The actual increase in risk can be reasonably estimated. In urban areas in daylight, 0.2% of accidents to cyclists are caused by motorists hitting lawful cyclists from behind. The maximum rate at which one lane of motor vehicles can pass a point is 2,000 per hour in heaviest freeway traffic. In urban conditions, where traffic signals and other delays break up the flow, the maximum rate at which motor vehicles can pass a cyclist must be less than 1,000 per hour.The average rate at which motor vehicles pass a cyclist is probably around 50 per hour. Therefore, the maximum cannot be more than 20 times the average. Therefore, the maximum increase in accidents produced by the maximum volume is 20 x 0.2% = 4%. The effect has no practical significance; many other effects are likely to be much more significant, such as the beneficial effect of good intersection channelization and traffic signals, and we know that the beneficial effect of knowing the proper cycling methods far outweighs this small increase in risk.

The definition of a road along which sight distance is inadequate is that on which motorists travel at speeds at which they cannot control their vehicles within the distance they can see ahead. "The sight distance is likely less than that needed for a motor vehicle operator to either change lane position or slow down to the bicyclist's speed." That is clearly violating the basic speed law. In those locations, "providing for bicycle operation to the right of the designated motor vehicle lane [by] a bike lane or shoulder" is what the manual calls the appropriate corrective action. In other words,for roads on which the majority of vehicles are traveling at unlawful speeds the appropriate measure is not to compel motorists to slow to the lawful speed, but to continue to allow it.

There we have the full statement of the scientific work that the FHWA presents as its justification for its 20-year-old bike-lane policy.The only new work is that drivers learn to compensate for the added dangers created by bike lanes, but both that and all the previous work demonstrate that bike lanes create more dangers than they prevent. The only words referring to safety are the weasel words about increased potential risk of high traffic volumes, which are debunked above. There is no scientific support for the idea that the FHWA designs reduce accidents to cyclists.

Wilkinson provides two justifications for his recommendations about types of cyclist.

The first justification is that "B/C" cyclists "prefer well-defined separation of bicycles and motor vehicles on arterial and collector streets (bike lanes or shoulders), or [to be] on separate bike paths." Wilkinson is saying here that the preferences of those who are ignorant should take precedence over the knowledge of those who know.

The second justification is that most cyclists (95% in Wilkinson's estimate) who are now ignorant of cycling in traffic will continue in their ignorance and "there will be more novice riders than advanced riders using the highway system."

Combining children with adult beginning cyclists is a politically potent maneuver that lacks intellectual content. Adults riding to their destinations (places where the parents of young children don't allow them to go alone) rarely use the same routes as children riding to elementary school or to their friends' houses, yet the FHWA says that routes primarily used by adults are to be designed according to the same standards that apply to children in elementary school. Our emotions about children on bicycles are being used to justify designing the general cycling transportation system as if for use by elementary-school children.

There is no evidence that the bike lanes specified for Group B/C riders make cycling safer or that they allow poorly-skilled cyclists to ride safely. The evidence from elsewhere is that bike lanes make cycling more dangerous and make it more difficult to learn. The only justification stated for the bike lanes that are specified is that they make B/C riders more comfortable. In other words, the FHWA is prepared to endanger poorly-skilled cyclists because, in their ignorance, they feel that the more dangerous facility is the safer. In this analysis there is no difference between group B adults and group C children. Bike lanes cannot make cycling safe for children.

If some program of attracting people to cycling, Wilkinson's or some other, succeeds in that effort, then the population will consist of people like me, who has been cycling for 56 years. It is a little stupid to think that I was a novice for more than 28 of those years. You can't have a population in any activity in which most participants are novices unless the average length of participation is much less than the time to learn the activity. People can learn how to ride properly in a few months, just as they can learn to drive a car. Nobody would be so stupid as to say that there are more novice drivers on the American roads than there are experienced drivers. The statement that when a considerable number of Americans ride for transportation there will be more novices than experienced cyclists is too obviously stupid to even be a calculated lie.

Certainly, there will be a changeover period. The shorter it is, the greater the proportion of novices at one time, but the shorter the time span. The longer it is (and I am not so optimistic as to expect a short changeover) the smaller the proportion of novices but the longer the period, but perhaps not proportionally long because there will be a greater proportion of experienced cyclists available to train the inexperienced. However, the highway system is a durable capital good with a long lifespan. Even roadway surfaces are expected to last 20 years. Sections of the system are expected to last twice that, at least, while the system as a whole is expected to last several hundred years. It is foolish to design the system in a way that is suitable for its novices because there may (the probability is not high) be many novices for one short period.


Wilkinson has made his recommendations on the basis of how ignorant people feel about bike lanes and bike paths ("B/C" cyclists "prefer" these facilities). Neither Wilkinson nor the FHWA makes any claim that bike lanes reduce accidents to cyclists, and they admit that bike lanes make the driving task more difficult. The policy is therefore entirely based on the cyclist-inferiority phobia and the desire of the motoring establishment to give overtaking motorists the right-of-way so they won't be delayed.

Wilkinson and the FHWA believe, recommend and presumably want the majority of an increasing number of cyclists to continue believing the cyclist-inferiority superstition and to remain "B" cyclists. While Wilkinson himself may not believe that having dumb, frightened cyclists is desirable as a benefit to motorists (the FHWA historically has had that opinion), he and the FHWA are so entwined by the cyclist-inferiority complex that they can't escape it. It dominates everything that they do, even when they think that they are doing good for cyclists. The facts are directly contrary. Since believing in these superstitions is the greatest single obstacle to reducing the accident rate for cyclists, and since the FHWA manual does nothing to reduce cyclists' accident rate, that manual constitutes a program for an increasing number of accidents to cyclists.

Return to: John Forester's Home Page                                             Up: Facilities