Cars: How did our worst idea become our only idea?
I saw a question on social media about "how did our worst idea become our only idea?" in response to car traffic.
Surprisingly, I know the answer to this, both practically and theoretically. Get comfortable and see what you think of my traffic manifesto:
Network theory says that the value of a network grows as a function of the nodes connected, generally on the order of a square law (the factor is argued, but not the general statement). This is known as Metcalfe's Law. We already collectively knew this for millennia, which is why paths, streets, and roads have grown to connect everything related to humans, and you hear stuff like "all roads lead to Rome" and such. The dawn of the telephone age formalized Metcalfe's Law, and the Internet has provided another data point, and recently Amazon and AirBnB have yet again. With any "iteration with preferential selection" situation, you end up with winner-take-all patterns, as more users add push for destinations to connect, and more destinations push more users to connect, and more connections expand the network.
As for human transport, not much differed if all were on foot, or some were on horses (note that everything that people complained about cars was once said about horses and carriages, which too often ran over poor kids with little repercussion). A carriage could go too fast for existing traffic, but get to more places conveniently, so all would naturally gravitate to horses if they could. But horses are expensive, so a secondary feedback loop prevented a horse-traffic takeover.
In the 1800's, railroads arose for the same reason. A steam train is fast compared to a horse, and much cheaper for an individual trip/load, so rail took over for longer-distant connectivity, and with later electrification streetcars came along also for local travel. Trains and trams have complex and expensive infrastructure, so it's hard to have rail as a door-to-door solution, hence ToD and DoT became valid concepts. Rail did not really compete with streets, other than to create dangerous at-grade crossing - the networks connect and overlap, but do not directly compete. Rail did compete with roads between towns, but generally rural right-of-way land was pretty cheap so dual networks evolved.
After 1875, a bunch of interesting things happened. Electrification brought along sewing machines and washing machines for the home, and completely changed the central power model of factories. That's and interesting story....but also at the same time came the bicycle. Most people today do not know that bicycles quietly took over the world after the Penny Farthing (and its coining of the term "took a header") gave way to the "safety bicycle". Cycling manufacture around the world exploded, and paved streets, which were already common in big European cities, grew in expanse. Bikes were cheap compared to a horse, and cleaner, and became popular with men going to newly distributed craft jobs driven by machine electrification and with women with new spare time due to sewing/washing machines.
This is where we see "iteration with preferential selection" playing its game -- for every trip, a person could decide what conveyance to use, and for each purchase, whether to buy a horse, a bike, a new pair of shoes, etc. Most people did not buy railcars (though some did), and many could not buy horses (though some did, and complained a LOT about cyclists taking "their" streets), but they could buy a bike. With each new bike, another person could easily reach more distant destinations, and with each new cyclist there was value to have housing and shops a little further out. At the same time, electric trams/streetcars were invented, and these too used street space in cities, but were relatively few in number so the time-space occupancy was small and speeds were limited to manage risk acceptably. Theory would say that soon enough, cycling and streetcars should take over, and cities would also expand. And they did.
Surprisingly, I know the answer to this, both practically and theoretically. Get comfortable and see what you think of my traffic manifesto:
Network theory says that the value of a network grows as a function of the nodes connected, generally on the order of a square law (the factor is argued, but not the general statement). This is known as Metcalfe's Law. We already collectively knew this for millennia, which is why paths, streets, and roads have grown to connect everything related to humans, and you hear stuff like "all roads lead to Rome" and such. The dawn of the telephone age formalized Metcalfe's Law, and the Internet has provided another data point, and recently Amazon and AirBnB have yet again. With any "iteration with preferential selection" situation, you end up with winner-take-all patterns, as more users add push for destinations to connect, and more destinations push more users to connect, and more connections expand the network.
As for human transport, not much differed if all were on foot, or some were on horses (note that everything that people complained about cars was once said about horses and carriages, which too often ran over poor kids with little repercussion). A carriage could go too fast for existing traffic, but get to more places conveniently, so all would naturally gravitate to horses if they could. But horses are expensive, so a secondary feedback loop prevented a horse-traffic takeover.
In the 1800's, railroads arose for the same reason. A steam train is fast compared to a horse, and much cheaper for an individual trip/load, so rail took over for longer-distant connectivity, and with later electrification streetcars came along also for local travel. Trains and trams have complex and expensive infrastructure, so it's hard to have rail as a door-to-door solution, hence ToD and DoT became valid concepts. Rail did not really compete with streets, other than to create dangerous at-grade crossing - the networks connect and overlap, but do not directly compete. Rail did compete with roads between towns, but generally rural right-of-way land was pretty cheap so dual networks evolved.
After 1875, a bunch of interesting things happened. Electrification brought along sewing machines and washing machines for the home, and completely changed the central power model of factories. That's and interesting story....but also at the same time came the bicycle. Most people today do not know that bicycles quietly took over the world after the Penny Farthing (and its coining of the term "took a header") gave way to the "safety bicycle". Cycling manufacture around the world exploded, and paved streets, which were already common in big European cities, grew in expanse. Bikes were cheap compared to a horse, and cleaner, and became popular with men going to newly distributed craft jobs driven by machine electrification and with women with new spare time due to sewing/washing machines.
This is where we see "iteration with preferential selection" playing its game -- for every trip, a person could decide what conveyance to use, and for each purchase, whether to buy a horse, a bike, a new pair of shoes, etc. Most people did not buy railcars (though some did), and many could not buy horses (though some did, and complained a LOT about cyclists taking "their" streets), but they could buy a bike. With each new bike, another person could easily reach more distant destinations, and with each new cyclist there was value to have housing and shops a little further out. At the same time, electric trams/streetcars were invented, and these too used street space in cities, but were relatively few in number so the time-space occupancy was small and speeds were limited to manage risk acceptably. Theory would say that soon enough, cycling and streetcars should take over, and cities would also expand. And they did.
So,
in essence, bicycles and streetcars could leverage the same network to
better advantage than previous transmission protocols, and yet coexist,
so the network adapted to match. Nobody notices the history of cycling
now, as bicycles are small and fit pretty well into a world with lots
of people, some horses, and a few carriages. I'm sure they had parking
issues, but this doesn't much show up in the "hard assets" we see as
streets have evolved. Streetcars are still remembered, given the
"streetcar suburb" housing layouts and continued existence of derelict
rails peeking through here and there.
Then along came cars. Cars offered the same value that horses enjoyed, and the same issues, only more-so. When there were only a few, they were owned by the wealthy, and the same issues were handled the same way: "don't get in my way!". Since as we've seen the network connectivity value is driven by network size, and that is determined by speed (given a static density), going fast was desirable. Early on, we had a few fast cars trying to go more places, and cars competed with bicycles and horses and then streetcars, but the secondary limit of price was still in place. Note that cars also competed with rail, where connectivity existed between towns.
As car prices dropped, the iteration and preferential selection mechanism did its thing, and as we know that for any trip one can choose whether to walk, bike, ride, or drive, and often the car was (and is) the fastest solution from A to B, so it gets picked.
But over time, as car volumes increased, the network itself had to grow, and the combination of demand for road space and parking plus the growing hazard to other traffic types first impact street usage, and then the roads, and finally the city layouts themselves. As with bikes and streetcars, higher speeds enable cities to grow at the edges, but the desire for connectivity is insatiable (Metcalfe's Law is even bigger than linear, remember!), so speeds must tend to go up as well. Unfortunately, the demand for speed and space presses back on the structure of cities, which responded by becoming less dense (since costs are lower at the edges) and more car-centric, as increasingly each trip choice would favor the car. Where streetcar suburbs tend to be dense and close-in, car suburbs tend to be sprawling and thus further out.
What limits car traffic? We have seen that this is actually known, with the Downs-Thomson Paradox -- it grows until another mode of connectivity is faster. This is precisely what the "iteration with preferential selection" networking model would say -- if another mode works better, it'll get picked. But now we know that cars, bikes, and peds (and rail) do not really share the same network without friction. Cars took over streets completely, so to use another mode a new network has to be built, and this requires explicit decisions that are NOT part of the daily iterative selection and preference process.
There is absolutely nothing new about this, as transit planners have been saying for a long time that headway frequency is critical, and yet the destination coverage is critical too. The square law of network theory says that transit will be pulled toward covering more destinations, but the preferential selection per trip says that transit will only be used when it is the fastest choice.
This also means that on-street buses, as in my city, will never successfully compete with cars. It is theoretically and practically impossible....the only point favoring on-street buses is cost, and even that requires public support. The only people who will use such buses are people who do not have cars, and that makes sense as a choice for them ONLY IF the cost is lower, AND their destinations are reachable by bus, AND the time is not too long. For others, a bus might make sense if a lethargic bus connection is a small inconvenience on the end of a faster trip using another mode. So, if you have a fast train into town and then a slow bus, you might still see the combination as better than a slow drive with parking hassles and expense.
So, what does this say:
Then along came cars. Cars offered the same value that horses enjoyed, and the same issues, only more-so. When there were only a few, they were owned by the wealthy, and the same issues were handled the same way: "don't get in my way!". Since as we've seen the network connectivity value is driven by network size, and that is determined by speed (given a static density), going fast was desirable. Early on, we had a few fast cars trying to go more places, and cars competed with bicycles and horses and then streetcars, but the secondary limit of price was still in place. Note that cars also competed with rail, where connectivity existed between towns.
As car prices dropped, the iteration and preferential selection mechanism did its thing, and as we know that for any trip one can choose whether to walk, bike, ride, or drive, and often the car was (and is) the fastest solution from A to B, so it gets picked.
But over time, as car volumes increased, the network itself had to grow, and the combination of demand for road space and parking plus the growing hazard to other traffic types first impact street usage, and then the roads, and finally the city layouts themselves. As with bikes and streetcars, higher speeds enable cities to grow at the edges, but the desire for connectivity is insatiable (Metcalfe's Law is even bigger than linear, remember!), so speeds must tend to go up as well. Unfortunately, the demand for speed and space presses back on the structure of cities, which responded by becoming less dense (since costs are lower at the edges) and more car-centric, as increasingly each trip choice would favor the car. Where streetcar suburbs tend to be dense and close-in, car suburbs tend to be sprawling and thus further out.
What limits car traffic? We have seen that this is actually known, with the Downs-Thomson Paradox -- it grows until another mode of connectivity is faster. This is precisely what the "iteration with preferential selection" networking model would say -- if another mode works better, it'll get picked. But now we know that cars, bikes, and peds (and rail) do not really share the same network without friction. Cars took over streets completely, so to use another mode a new network has to be built, and this requires explicit decisions that are NOT part of the daily iterative selection and preference process.
There is absolutely nothing new about this, as transit planners have been saying for a long time that headway frequency is critical, and yet the destination coverage is critical too. The square law of network theory says that transit will be pulled toward covering more destinations, but the preferential selection per trip says that transit will only be used when it is the fastest choice.
This also means that on-street buses, as in my city, will never successfully compete with cars. It is theoretically and practically impossible....the only point favoring on-street buses is cost, and even that requires public support. The only people who will use such buses are people who do not have cars, and that makes sense as a choice for them ONLY IF the cost is lower, AND their destinations are reachable by bus, AND the time is not too long. For others, a bus might make sense if a lethargic bus connection is a small inconvenience on the end of a faster trip using another mode. So, if you have a fast train into town and then a slow bus, you might still see the combination as better than a slow drive with parking hassles and expense.
So, what does this say:
- Alternative choices about transit architecture can NEVER be made by auto interests, including drivers themselves. Once you have paid for a car, the incremental cost per trip is small, and it will self-justify the auto -- it's a stable and durable state.
- On-street options, like ordinary buses and bikes, will not be acceptable trip options UNLESS they are separated from other traffic via dedicated lanes. Buses are impacted by the same delays as cars, yet with fewer destinations available and more stops to slow them. BRT will only attract drivers if given dedicated lanes and priority.
- Parking is part of the problem in that convenient, free parking favors cars in multiple ways. If parking costs more, autos are less affordable as a per-trip option. If parking is hard to find, it adds to the effective trip time, and impacts the trip choice trade-off IF other options are available.
- If
all traffic is by car and bus, car traffic will tend to increase until
the travel cost (time and money) exceeds the trip value, unless there is
a point where another mode makes sense. In some cities without transit
this happens when walking is faster than traffic-jammed streets, yet
even for this situation cars hold an edge by making walking physically
hazardous, noxious, and long due to sprawl. Without options besides cars, the network will favor cars and
nothing else, not even human life in general. You'll be able to go
anywhere in theory, but most trips won't be worth the time and hassle,
and crashes and pollution will continue to be major issues for humanity.
- Thus,
separate network infrastructure is critical to enable reasonable
choices along with decent living. With convenient options for transit
and cycling, or even walking, cities can be denser and more connected,
AND car travel will work better too. Oddball trip connections can still
be made by car, while most routine trips can more easily and
effectively be done using another mode. Metcalfe's Law and the
Downs-Thomson Paradox will work together to expand useful connectivity,
while our explicit choices about transit network investments will reign
in the most undesirable aspects of each option.
But that's an engineered solution, politically and physically. What can ordinary people do, cumulatively, that ordinary people can cluster around and amplify, that will inadvertently solve the problem of roads?
ReplyDelete