Motivation is an interesting thing. As a manager I was always interested in how best to motivate employees to achieve what was most important for our company. Many people that I hired were motivated simply to do a great job and for these folks the best thing to do was to give them what they needed and get out of the way. The result would consistently be phenomenal.
This hasn’t seemed to be the case with traffic engineering in the U.S. The result hasn’t been phenomenal. It’s been a failure. In fact, we have the the most dangerous road system of all developed countries and, thanks to gobs of lanes and high speed traffic, some of the most uninviting places.
In The Goal Eli Goldratt makes a point of the problems that are caused by focusing too much on one single metric—efficiency. This may also be key in the failure of traffic engineering in the U.S., the seemingly single-minded focus on Level Of Service, or LOS.
One problem with LOS is that it is not a measure of level of service at all but only of level of delay for motor vehicles.
LOS | Delay (Seconds) |
---|---|
A | < 10 Seconds |
B | 10 - 20 |
C | 20 - 35 |
D | 35 - 55 |
E | 55 - 80 |
F | > 80 Seconds |
MnDOT and other agencies talk a lot about safety and about serving the needs of ‘all users’ but what apparently happens when things get down to the drawing board is the only thing left is this lone number and that is what drives the design. Safety is apparently no longer so critical. Nor are the needs of all users. Nor is the impact on people who live, shop, or work nearby. Only LOS.
This certainly makes life easier for traffic engineers. And more dangerous and unpleasant for the rest of us.
People be damned, just get the cars through.
True LOS
Let’s make LOS a better measure and align it with our goals.
Along with good flow of traffic, we want to reduce deaths, injuries and the negative impact of traffic on people who live, work or shop near our roadways. We’d like to better serve all users including pedestrians, disabled, and bicycles. Fortunately these are mostly mutually beneficial, for instance, increasing the mode share of walking and bicycling reduces motor vehicle traffic, noise, pollution, and danger. Less motor traffic decreases delay at junctions.
First, let’s include delay for all users, not just cars.
On the safety side let’s add four measures; motor vehicle (MV) deaths per capita, MV injuries, Pedestrian/Disabled/Bicycle (PDB) deaths and PDB injuries[1].
We also want to make things more safe, comfortable and inviting for active transportation — people walking, riding bicycles, or disabled — so let’s include a mode share element. This will be measured for each junction whenever possible and will be a reduction in points (improvement). Each percent improvement will be worth 3 points.
Some Motivation
Since traffic engineers don’t seem to be motivated by doing a good job and creating a great roadway system, perhaps we can motivate them with money.
Note: Though I would lump most traffic engineers I’ve encountered in to this group, there are a growing number who are fighting hard to bring the profession out of it’s antiquated 1950’s belief system that has given us the world leading death rates we have today.
Similar to pay in the private sector, let’s set the pay of traffic engineers so that 30% of their current salary is ‘at risk’ and based on the LOS that they achieve. For easy math let’s assume they currently make $100k so now $30k will depend on how well they do their job.
Since they don’t seem to like too much change let’s keep the basic LOS scoring system where the goal is as low a score as possible. So, a LOS of zero is ideal and about 25 is still their initial target and achieving that will pay 100% of their target pay. For each point better (lower score) they will receive an additional 3% and for each point worse (higher score) they will lose 2%.
LOS Attainment | Payout % |
---|---|
1 | 172% |
... | |
23 | 106% |
24 | 103% |
25 | 100% |
26 | 98% |
27 | 96% |
... | |
75 | 0% |
Since we are so far behind other developed nations and it will take some time to catch up, it wouldn’t be fair to engineers to dump this expectation on them all at once. Let’s phase it in over 5 years at 6% per year. So, in 2015 6% of their salary will be based on LOS achievement, in 2016 12%, etc.
Example:

County 96 and Hodgson Rd in Shoreview, MN. Lower right is a grade school. Other corners are a mix of office and retail. There are two senior housing coops just to the north of the image as well as a middle school. This area gets fairly high pedestrian, bicycle, and disabled traffic for a suburb thanks to the generally very good MUPs in the area.
County 96 and Hodgson Road in Shoreview.
Current LOS is B with an average delay of 19 seconds[2]. Average delay for bicycle riders and pedestrians however is about 64 seconds which gives us an overall average of 41.5[3].
Add 14.42 points for motor vehicle deaths in Ramsey County, 17.64 for pedestrian/disabled/bicycle deaths, 12.77 for MV injuries and 14.80 for PDB injuries which all gets us to 101.13.
The mode share for this specific junction is a guesstimated 0.5% for pedestrians and 1% for bicycle riders so we get a 4.5 point reduction leaving us at 96.63 points for this junction.
County 96 & Hodgson Rd (Current)
96 & Hodgson (Current): | Value | Weighting | Points |
---|---|---|---|
Delay: MV | 19 | ||
Delay: Ped/Bicycle/Disabled | 64 | Avg | 41.5 |
MV Deaths (per 100k population): | 2.88 | 5 | 14.4 |
Ped/Bicycle/Disabled Deaths: | 0.71 | 25 | 17.75 |
MV Injuries | 638.47 | 0.02 | 12.7694 |
Ped/Bicycle/Disabled Injuries: | 59.21 | 0.25 | 14.8025 |
101.2219 | |||
Pedestrian Mode Share: | 0.005 | 3 | -1.5 |
Bicycle Mode Share: | 0.01 | 3 | -3 |
Total Score: | 96.7219 |
So, what’s a traffic engineer to do?
Their initial goal is to get this and other junctions down to 25 or lower. Neither increases in safety nor mode share alone will likely get them there, but modest improvements in each will.

One slip lane plus 7 other lanes to cross this thing. A safe left turn for pedestrians, disabled, or bicycle riders using crossings and signals can take up to 4 minutes. I ran this junction, including MNDOT traffic volumes, by a Dutch engineer. He said that Hodgson Rd (the north-south road) would likely never be more than two lanes at any point (it’s 4 to 6 lanes here). Ramsey Cty 96 (east-west) is borderline but would likely be two lanes if this were a roundabout (which he said could be either a single or two lane roundabout) though might be four. They would never have the slip and turn lanes we have.
As a department their first focus will be on decreasing deaths and injuries across the county since this will effectively improve LOS at every junction. They’ll lower some speeds[4], begin eliminating rights-on-red and slip lanes, and introduce a lot more roundabouts. We’ll likely see fewer lanes in general and many fewer turn lanes (especially at junctions with high numbers of people).
I wonder if they’ll discover that narrower driving lanes and roundabouts decrease distracted driving and improve safety since they require a bit more attention than our wider lanes and intersections? Whatever they do, we’ll have the traffic engineers attention and now they’re focusing on our safety rather than just speeding us through intersections.
An initial target of a 33% reduction in both deaths and injuries would seem attainable (we’d need a 59% reduction to be average for all OECD nations, but that might be asking a bit much). This gets us to 61.95 and more importantly saves about 140 people from being killed by drivers each year.
I wonder who these 140 will be? Will you or I know one of them?
Increasing mode share may be the easiest. Make this and other nearby junctions very safe, comfortable, and efficient. Make walking or riding a bicycle just as good as or a better alternative than driving for short trips. With much of traffic through this junction being somewhat local (rather than long distance commuting) a mode share of 5% for pedestrians and 6% for bicycles and disabled should be reasonable and now we’re at 33.45.
Let’s not forget the mutual benefits of most of these changes. Less MV traffic, shorter light cycles, maybe a simultaneous green for people riding bicycles, and both MV and PDB delay have decreased to 18 seconds.
The result of all of this is a score of 24.95, a 103% payout, and thus a well deserved $3,000 bonus (before taxes) for our traffic engineers.
A bit more reduction in injuries or deaths gets us another point and a 106% payout.
——-
If you wonder how we’ll pay for these bonuses, the savings for Ramsey County from fewer crashes alone will more than cover the costs[5]. And that’s only the beginning of the monetary benefits not to mention all of the other benefits.
Fair? It’s not like we’ll be asking our traffic engineers to be as good as their peers elsewhere, but only about half as good. Is that asking too much?
####
[1] To guard against uncontrollable fluctuations one year to the next these will be three year running averages. We’ll use deaths and injuries per capita instead of per vehicle mile travelled because our overall goal is not just safer roadways but safer communities. This will help keep traffic engineers and planners focused on their role in the big picture not on one disconnected element. For instants, they will become interested in city planning elements that reduce the need to drive as much or as far. These will be calculated on a county-wide basis.
[2] The numbers used in this example are as accurate as I’ve been able to quickly obtain. Highly accurate numbers are not critical for our example purposes.
[3] Note that these are not necessarily comparable since motor vehicles often go through without even slowing while pedestrians, bicycle riders, and disabled must always stop to push a button and wait. A MV not slowing gets a delay of 0 while a bicycle rider who gets an immediate crossing gets a 1. However, the delay for the bicycle rider is actually much greater than this one point indicates.
[4] Note I didn’t say speed limits but speeds. The need will be to reduce actual speeds not just change speed limits.
[5] Costs of police, fire, and other services for each crash as well as damage to county property.
If we’re going to start increasing safety by deliberately ruining LOS for cars, why not lower the speed limit to 5 mph and require motorists to stop every 100 feet and wave a safety flare? I’d disagree with the entire premise of the article. Engineers do care about safety, but it’s done with benefit/cost analysis that also factors in time savings and such.
Well the current norm is to increase LOS by deliberately ruining safety for everyone. Why not raise the speed limit to 50 MPH and require pedestrians to wave safety flags and/or hit beg buttons? Oh, we already do that?
Walker does care about LOS, but it’s done with cost/benefit analysis that more heavily weights peoples lives over time savings and such.
Is it really an either/or issue? How is it that EU countries have 1/2 to 1/4 the road fatalities we do (per capita or per VMT) and without 5 mph speed limits and stops every 100 feet?
At the risk of pure speculation, how do those countries compare with respect to things the like penetration of car ownership/use, age of drivers (young, primarily, but old too maybe), cost of ownership, insurance and liability? Perhaps driving is limited to a set of drivers that is inherently less dangerous?
And, of course, vehicle size and power tends toward the less destructive as well.
Speculation is good. That’s where most great ideas come from. Off the top of my head… Most countries don’t allow you to drive until you’re 18 which may produce a safer set of drivers. You also don’t see nearly as many elderly drivers. Cost of initial and ongoing ownership is quite a bit higher. I believe both vehicle ownership and VKT are about 20% lower per capita (much lower in northern Europe not much lower in southern).
What miles they do drive are a bit higher ratio of motorway to urban which is safer. They are more likely to walk, bike, or transit for local trips than we are and they are much more likely to drive for summer holidays than we are.
These though don’t account for their much lower fatalities per capita (see the graph in the sidebar here: https://streets.mn/2014/07/01/bicycling-relative-safe/)
Their fatalities per VKT are still much lower (see other charts on that link) and are much lower in every age group indicating that their roads themselves or how they use them are much safer.
In the end though it’s still the overall fatalities per capita that matters the most as we’re designing cities and transportation (though fatalities per VKT matters most if I have to choose between two modes of transportation for the same trip).
Monte, I think we all believe you when you say engineers care about safety. I think the question raised by this article is, how do they factor safety, quality of life, other transportation modes, etc into their cost-benefit analysis? It appears to me that Walker is trying to propose a way to more explicitly include a variety of factors, and not just focus on the time it takes a car to travel an area. Anyway, any insights on the engineer’s process?
“Engineers do care about safety, but it’s done with benefit/cost analysis that also factors in time savings and such.” Really? Have they failed that miserably? We have the most dangerous road system of all developed countries. More people are killed per capita and per mile driven in the U.S. than any other developed country.
This statement doesn’t seem right. Where are you getting your data? I did some digging (Wikipedia) and found this report. http://internationaltransportforum.org/Pub/pdf/14IrtadReport.pdf
U.S. is in the lower two-thirds of developed countries in fatalities per billion VKT but ahead of Belgium, Japan(!), New Zealand but definitely not last. Also, in your other article you showed that MN’s fatality rate is about 2/3 of the national average. So, applying that math, MN’s fatality rate per bilion VKT is about that of Germany’s (~5 vs. 4.9). So, yes, MN transportation engineers are doing as bad a job as those notoriously inept German engineers.
You also apparently haven’t heard of Minnesota’s Towards Zero Deaths program which focuses on the four E’s (engineering, education, enforcement, EMS) because they all impact traffic safety. I would guess that most European countries have far lower rates of drunk driving for example. There’s also policy. I would guess that most European countries have stricter laws regarding impaired driving, motorcycle helmet use, seat belt use. I could go on.
John, I think you are correct and I misspoke. We (the U.S.) have the most dangerous road system per capita but not quite the worst per VKT. More later…
This is a great example of why I should refrain from comments and stick to stuff where I (usually) take time to review all my numbers and statements. On the other hand, this is a good and important discussion.
The IRTAD report (great org btw) indicates that, at least based on OECD countries, we’re no longer the worst in per capita fatalities either, with Korea and Chile worse than the U.S. in 2012. So we’re 32nd out of 34.
The good news is that we are improving. We still have a long way to go however. Our goal needs to be achieving what those in the top third achieve, not being in the bottom third.
In any case, please don’t let my errant comment detract from the core of the issue. Not being the worst isn’t much to crow about.
I’ve had two longtime concerns about traffic engineering studies:
1. There is the assumption that volumes will continue to rise everywhere, so let’s build future capacity. This has often been done in the center cities when in fact volumes have remained stable or even declined over the decades. This has resulted in destructive road widening, such as Lowry Avenue N.
2. The studies don’t take into account that traffic looks for alternate routes when confronted with congestion. That reduces the increases in volumes.
Detouring congestion goes mainly on arterial streets is the funny thing. Going down sidestreets is relatively rare. Often you get more complaints of people speeding up and more traffic and I’ve actually found less traffic and slower speeds.
Traffic looking for alternative routes isn’t necessarily desirable though, for example if you don’t build turn lanes on major streets so traffic starts to back up, cars will divert off into residential areas trying to avoid it. I’d have to agree that there’s a lot of fantasy in long term predictions, though. Expanding the Winona bridge to four lanes was driven in part by studies predicting a need over twenty years from now. The thing to do would have been to build a two lane bridge now while saving the old historic bridge for non-motorized traffic meaning it would last a lot longer, then build another two lane bridge when (and if) the forecast traffic ever materializes.
Just a thought: the notion that “residential streets” (ignoring, of course, that many more people per foot-width of street live along urban arterials like Lyndale, Hennepin, etc than many residential areas) should be protected from traffic flows, even marginal increases, is entirely bizarre to me. Here we have streets that are paved with expensive materials to support cars, but only intended for extremely low volumes of traffic just because it makes the lower-density residents happy? This notion results in busy arterials that act as barriers to crossing by design (loooong light cycles or none at all which may mean you never cross), funneling more vehicles onto the arterials to get around. Freeways are even worse as they cut off crossing points entirely. In MSP, most of the barriers that force auto traffic onto arterials are due to these man-made barriers (as opposed to lakes, rivers, etc one typically thinks of that increase costs of connections).
I would be completely fine with my street (Fremont Ave S, between 35th and 36th) tripling its traffic counts as long as they were calmed (20 mph). I would be willing to trade the 4-way stop sign at 35th (& beyond) for a small neighborhood roundabout that kept cars moving slowly (negating the time lost from the 20 mph design speed). If we were able to divert even 1/3 of arterial traffic to ‘residential’ streets, would we need as many lanes on the arterials themselves? Would biking (or dedicated space for transit) become possible? Would mode shifts from improved services/infrastructure further change the equation?
I think that deserves a post of its own. You’re definitely on to something by pushing back on the idea that the purpose of “residential streets” should essentially be private streets only for the residents who live on them. In general, I agree that increases in “cut through” traffic are ok as long as it is calmed. Obviously there are situations where the “cut through” traffic volume or speed gets out of hand and needs to be dealt with by infrastructure changes.
At what point though does this increased traffic make the street less inviting and less comfortable for average pedestrians, disabled, and bicycle riders? If they avoid it because they feel unsafe then what will happen to MV speeds?
If we spread traffic out like this do we create a need for massively more miles of segregated bicycle facilities than if we concentrate heavy traffic and only have to build segregated facilities along collectors and arterials?
I don’t think taking N-S streets like Emerson, Dupont, Fremont, etc that see maybe 500 cars/day each and assuming they can handle more like 1,500 would make them un-bikeable or less appealing to walkers. Most of these streets have boulevards with trees separating the sidewalk from the street, and cars parked along the way. Certainly bulb outs that denote intersections as pedestrian-first would help keep traffic calmed & enhance comfort.
Obviously, this is just a thought experiment. As it stands, 94, 394, 35W, etc (my S Minneapolis lens showing through) limits these streets from having connectivity to the rest of the city without funneling into the arterials at some point, so it’s mostly moot. But I’m just challenging the assumption that “neighborhood streets” shouldn’t/couldn’t handle slightly more vehicles if the tradeoff means our arterials (where transit runs and we’d like to carve space out for cycle tracks) take less (allowing those improvements, both technically and politically).
Would you send your 8-year-old daughter out to ride a mile to her friend’s house on the streets as you propose them?
Yes. I guess I don’t see a huge difference between 500 cars and 1,500 cars a day (how many cars per minute), especially if the streets are actually calmed. And the trade off would be worth it if it allowed arterials like Hennepin/Lyndale/etc to gain protected cycle tracks and become more crossable.
One of the likely problems is that people who live on a street and use it only for access to their home are much more likely to drive slowly and carefully. People just using it as a rat-run are less likely to drive slowly and carefully.
Keeping streets to shorter segments for cars (pseudo woonerfing?) should be more effective in speed control than roundabouts and other calming measures (though they are still valid methods). This would interrupt your ability to use these as through streets.
Also, theoretically and statistically anyway, an increase in traffic increases the risk by at least that much. And decreases the comfort for PDB’s. And, 1500 cars per day is likely to have at least three times as many speeders and errant drivers as 500 though likely many more than three times as many since the added traffic are rat runners. And at least three times as many bass thumpers.
On the other hand, all of these cars added to an arterial (or not taken away from it) make very little difference to users of a properly designed segregated cycleway along that arterial.
And finally, there is the hope that by increasing trips made by walking or bicycling we will decrease the number of cars on the arterials in the first place. I’d love to know how many of the cars along Grand Ave in St Paul or Central in Minneapolis are making less than a 5 mile round trip.
Well I think we are all agreed that if people bike more things would be better, so there’s not much argument there.
But on the subject of keeping all traffic on arterials (where people live too) vs sending a few more down each side street, I’d have to agree with Alex and Janne.
I know it’ll be tough for those homeowners to have to hear 6 more cars with bass throughout the day, but that is also 6 less cars with bass that the people living on the arterial will hear.
The problem though, as The Netherlands, Denmark, and elsewhere experienced, is that when you do that you turn a street (or numerous) that can be safely and comfortably shared by MV’s and PDB’s in to one that requires segregated bicycle facilities for PDB’s to be safe and comfortable.
The extra traffic makes little relative difference on arterials but a very huge difference on otherwise calm residential streets. So instead of needing a segregated path along just one road, you now need it along the arterial and numerous side streets. Depending on, as Joseph mentioned, how many actually switch from the arterial to the residential streets.
If I remember I’ll pull out my CROW manual tonight and see what they determined was the maximum speed and volume for comfortable sharing and where they determined it starts to have a significant negative impact on PDB’s.
A bit from the CROW manual:
Anytime auto traffic is greater than 30kph (18 mph) a separate cycle facility (cycletrack, parallel cycle road, or side path) is required. Cycle lanes are no longer recommended.
Table: Auto/Bike/Disabled traffic can be combined up to a max of 5,000 pcu’s per day. However, a cycle track is strongly advised for over 4,000 pcu’s and recommended for over 2,000 pcu’s. These are for all traffic (MV/bike/disabled) in both directions.
The discussion then makes it clear that these are absolute maximums assuming ideal conditions for fairly short residential roads. For example, if actual speeds are greater than 18 mph then a separate facility is called for. Speed and volume is usually controlled by limiting the length of the streets or using opposing one-ways to prevent non-local thru traffic. Also, parking is alternated to induce slowing (e.g., creating chicanes).
A narrow road profile that physically prevents automobiles from passing bicycle riders is recommended.
There is also discussion of the travelable length of a street (between T junctions, auto barriers, or opposing one-ways), traffic volume, and auto-bike encounters that is to be considered. The gist of it being to limit the number of encounters.
Elsewhere there was discussion of ratios with caution about combining traffic if MV’s make up more than some percent of vehicles. The higher the volume the lower the percent of MV’s that should be allowed. I think the maximum was something like 70% MV’s.
All of this should be viewed in context. The CROW manual calls for and Dutch traffic engineers design for the most vulnerable so anytime there is a question then side with making PDB’s safe and comfortable.
I want to thank Alex for recognizing the (in the core cities) ARE residential streets. There are many, many apartment buildings along Lyndale (North and South), Franklin, Hennepin, Broadway, Park, Portland, Central, Lowry, Glenwood, Plymouth, Washington, etc. etc.
Now, I’m sure the people living in those buildings were fully aware that they were moving onto streets with lots of traffic when they chose those homes, but that doesn’t negate the importance of protecting and better yet improving THEIR quality of life.
Surprisingly traffic doesn’t detour onto these side streets as much as many people think it does.
The author said: “An initial target of a 33% reduction in both deaths and injuries would seem attainable (we’d need a 59% reduction to be average for all OECD nations, but that might be asking a bit much).”
I’m all for traffic calming, building streets and not stroads, prioritizing pedestrians and bikes on streets, etc. But even though this is vital to creating productive places where people want to live, we need to be realistic about its impact on traffic deaths and injuries. In 2012, 54 percent of fatalities were in rural areas, according to NHTSA. This is the case even though only about 20 percent of the population lives in rural areas, according to the Census Bureau.
Any serious attempt to cut traffic fatalities has to account for the disproportionate amount that rural roads contribute to the problem. These roads are not good candidates for the street calming measures described because they do not have the same purpose.
Again, I’m not saying don’t aim for 8-80 cities. I’m saying do it because it’s a good idea to build cities to human scale. If road safety is your main goal, though, it is probably best to focus on rural improvements first because those roads are the most dangerous both proportionally and in absolute terms.
In the picture of Hwy 96 and Hodgson Road, the lower left corner building with the red orange roof is a senior apartment building. It is where my parents live. There have been many close calls when the apartment dwellers have tried crossing Hwy 96 to get to the Walgreens in the upper right area or to what used to be a Rainbow (white roof building in upper left corner). The Rainbow is closed but hopefully another store will move in there.
In addition, that corner gets a lot of bike/ped traffice from the Ramsey County regional park trail system which intersects the Hwy 96 sidewalks just off the picture to the lower left.
Interesting article. Apparently “level of service” was not the key consideration when Minneapolis and Hennepin County decided to widen sidewalks along Cedar Avenue on the West Bank and restrict automotive traffic to one lane in each direction. I would like to know how the project really originated. Supposedly pedestrian needs were paramount (certainly not bicyclists’) but the City stubbornly refuses to install an illuminated red no right turn arrow on the northwestern corner of Cedar and Riverside, allegedly the most dangerous interesection in Minneapolis for pedestrians. Probably several hundred motorists make improper right there every day; I’ve come close to getting hit as I crossed Riverside.
I would really like to see a fuller back and forth about whether we should encourage arterial streets. From posts I’ve read by Walker it seems definitely yes, but I’ve read other posts that are all about completing the grid and decreasing these huge barriers.
In a brief conversation with a friend I concluded that we should go strong with arterials but just make them really well so that they are completely accessible and traversable. But I’m really not so sure.
Payne Ave on the East Side of St. Paul is a strong commercial street. It has half the traffic of parallel Arcade (US-61), which has half the traffic of Maryland Ave (CR-31). It is a main street that is not a functional through street, as it terminates early on the north and south while parallel streets keep going. Thus, it’s not a good arterial street but a good, local main street.
I currently feel boxed in by Arcade and Maryland where I live near Lake Phalen. I don’t let my 5 year old bike on the residential streets as they are (not woonerfs), and a little more traffic to make my neighborhood better connected would not bother me. I am leaning against arterials, generally.
Pingback: Sunday Summary – September 28, 2014 | streets.mn
Here’s an interesting related article: http://www.citylab.com/commute/2014/07/transit-projects-are-about-to-get-much-much-easier-in-california/374049/