One of the interesting and influential figures I met recently in Paris was Professor Donald Shoup from the University of California, Los Angeles - I have a great shot of him iding a (rented) Velib bike near a well-known Parisian tourist attraction. Shoup is one of the world’s experts on the economics of parking. This sounds like a dry topic but it isn’t – parking practices are, in fact, a significant contributor to the unpaid, social costs of motoring. I reviewed Professor Shoup’s The High Cost of Free Parking here. The average US parking spot costs more to provide than the typical car occupying it and underpriced parking is a significant cause of traffic congestion. The gist of Shoup’s argument is that anyone should be able to park anywhere at any time by paying a high-enough fee and that this would reduce excessive traffic partly by reducing socially-destructive search efforts to find a parking spot. One should set supply=demand in the parking market - allowing for entry and exit from spots you will do this if a 15% vacancy level among spots is targeted. I think a dozen quality PhDs in economics could be constructed in Australia on themes developed in Shoup’s book and all would have greater social payoffs than the current batch of mindless atheoretical, time series macroeconomic studies or the surfeit of ‘sophisticated’, useless game theory projects we are producing.
Free-parking in Australia is less on an issue than it is in the US – parking is much more typically charged for here at least in city areas – and free-parking provided by employers is subject to fringe benefits tax which prevents some of the worst US distortions. A more significant issue in Australia is that on-street parking is typically much cheaper than off-street parking so that people create congestion and social waste by searching around for on-street spots. This is a problem in Sydney and Melbourne. Underpricing on-street pricing is normally accompanied by inefficient forms of rationing such as limitations on the time you can spend at a metered spot.
I wrote a review of Australian parking policy here.
I am generally opposed to tacking environmental policies onto congestion or parking charges. That is not to say that these externalities do not exist – they do – but there are much better tax bases to capture these effects such as excises on fuel. These excises are more closely linked to the generation of pollution externalities, such as CO2 emissions, than are congestion conditions. A petrol excise targets anti-social, four-wheel drive gas-guzzlers as well as cars driving under congested conditions and, of course, such a tax should be part of a comprehensive effort to target greenhouse gas emissions (GHG). That is not to say that controlling things such as congestion will not reduce emissions since, by reducing periods of waiting-in-traffic and reducing distances travelled it will.
Professor Shoup sent me this article to me. It explains how congestion pricing can reduce GHG emissions by increasing average speeds and smoothing the traffic flow. Congestion reduces fuel efficiency and increases GHG emissions because it slows traffic and causes drivers to brake and accelerate frequently. The article estimates how congestion pricing can reduce GHG emissions in typical Los Angeles traffic. Increasing traffic speeds reduces emissions by 7-12% , and smoothing traffic flows reduces emissions by another 7-12%. The article comes from the excellent Access which contains various interesting ideas on transport.
Shoup likes parking policies as means of congestion control since they are both effective and cheap relative to rather expensive road pricing. On the other hand telematic devices which are useful for measuring and addressing congestion and other externalities can also be used to advise citizens where parking spots are available and facilitate electronic payments for such charges. In the US Shoup is also much concerned with employer subsidies for parking. The effects of existing parking subsidies are larger than the proposed congestion tolls and forcing employers to offer a cash alternative to a parking subsidy rids urban systems of this source of the distortion. This is the good applied microeconomics that my heart warms to. But in Australia we still have $9 billion in congestion costs in our large cities annually despite relatively hefty parking charges and the fact that employer-provided parking is taxed as a normal income benefit. I think congestion in Australian cities is much more than a parking issue.
[...] the complaints fly every time parking charges are introduced or increased on council owned land.Harry Clarke has a better considered and informed idea about the economics of parking than yours truly, but here [...]
Harry,
I hope you don’t mean your own PhD students by this!:
“”"”"”"”I think a dozen quality PhDs in economics could be constructed in Australia on themes developed in Shoup’s book and all would have greater social payoffs than the current batch of mindless atheoretical, time series macroeconomic studies or the surfeit of ’sophisticated’, useless game theory projects we are producing.”"”"”
I don’t have any PhD students doing mindless time series or game theory work.
See Streetline Networks, a San Francisco company for which Don Shoup is an advisor (disclosure: as am I). They provide technology infrastructure to be able to implement some of these ideas.
John, I think you origionally raised this with me here. At the Paris meetings a common theme of congestion pricing via telematics was to give drivers a bonus – to show that the technology served them as well as charged them a price. Telematic devices should be able to tell you where vacant parking spots are and should enable you to pay your parking fees remotely – this would make sense if you removed constraints on length of stay in a spot and simply charged a market-clearing charge – if you are delayed you activate a button on your device. Or maybe you can set it up so that charges are taken automatic
HC:
yes:
See Streetline on this, and the 3-minute video, and Streetline services relevant to this.
The ideal endgoal would certainly be:
1) As a driver, I tell my care where I want to go, it consults the network, and offers me a menu of available spots on the car’s electronic map, including pricing estimates. Maybe it tells me about ones that seem likley to be vacated, if I’m a ways away. Maybe I can “reserve” one. Pricing is dynamic, but I should be able to find out any time from a handheld what it’s costing me.
2) I drive and park, and a parking meter reads the car’s transponder and charges me for the time as I use it. (I.e., the SF Bay Area uses FasTrak for bridges, for example.
BUT: it is impossible to just magically jump to that, because to work, it needs:
- physical infrastructure (sensors, smart meters)
- network infrastructure
- smart vehicles
Hence, this becomes a strong “insertion strategy” issue, in which each step you take has to be be useful and cost-effective for somebody, without requiring *everybody* to do something. For example, people would scream if the only way to get over the Golden Gate were to have FasTrak … so of of course some lanes still take money.
In the parking case, there’s a whole sequence of possible approaches, depending on the city’s preferences, existing infrastructure, etc. One can:
0) No practical solution can depend on every car having some device in it; it takes too long for the installed base to move.
1) Just put sensors down in a “2-hour parking” area, that doesn’t even have meters, and any parking control officers can do what they do now, but without hassle of chalking tires, and with *their* route-sheets better optimized, or even better, handhelds that tell them “3 cars an hour over on next block.”
2) Put sensors in simple metered spaces.
3) Put in smart meters as well, as part of the same wireless mesh.
Then, depending on how smart the meters get about charging, and what they can expect from the cars, and what kinds of sensors they have, you can start approaching the end goal.
Of course, even step 1, with the right network software and support for smartphones, helps drivers find parking. The really key step is starting top build the wireless mesh physical infrastructure out there. Also, once you do that, the mesh turns out to be useful for other things, the same way as the Internet gets used for applications for which no one would ever build the Internet.
For examples, fire hydrants need to be pressure-checked periodically. A fire hydrant with a pressure sensor could easily wake up once a day, and report its pressure to the street mesh, miniscule data.