- Tolling and P3s boosted in Obama transportation bill
- Colorado dust-up over toll concessions
- Progress and questions on autonomous vehicles
- Express bus and express lanes for Tappan Zee Bridge
- Toll interoperability via smartphones
- Upcoming Conferences
- News Notes
- Quotable Quotes
Tolling and P3s Get Unexpected Boost in Administration Bill
The Administration’s proposed surface transportation bill, released the last week of April, generally followed the outlines that were expected: a 37% increase in funding over four years, supposedly paid for out of a major tax reform that is nowhere in sight, plus much larger increases for everything other than highways. But rather than repeat the criticisms in my March issue article, I want to focus here on some good news for toll financing and long-term public-private partnership (P3) concessions.
First, the bill would continue and improve two critically important financing tools that have been used in most of the large P3 concession projects financed over the past decade. It would continue the TIFIA credit-support program at the increased MAP-21 levels, while beefing up its professional staff to address recent bottlenecks in project approval. And it would increase by $4 billion the current $15 billion cap on tax-exempt revenue bonds (private activity bonds-PABs) that provide a more-level financial playing field for P3 projects in relation to government-run projects.
And to my pleasant surprise, the bill would essentially mainstream the existing three-state pilot program that allows the use of toll financing to reconstruct aging Interstate highways. That’s critically important, despite the lack so far of political consensus being reached to actually use toll finance in any of the three states that hold slots in the pilot program (Missouri, North Carolina, and Virginia). Because getting to yes on such a major change will be difficult, allowing all states to attempt this increases the odds that at least one pathfinder state will come up with a win-win approach that can then be a model for other states.
And not limiting each state to a single Interstate corridor also makes sense. What happened in North Carolina, which selected I-95 for toll-financed reconstruction, was that people who lived along I-95 said “Why us? How come the folks in Charlotte won’t have to pay tolls on I-77 but we over here will?” With the single-project limit removed, a responsible state DOT can develop and explain a 20-year plan to reconstruct all of its Interstates, based on age and widening needs. That way, everyone will know that their Interstate is in the queue for refurbishment, plus widening, if needed.
Unfortunately, the Administration’s tolling permission is not constrained enough to make the new tolls a pure user fee, similar to an electric bill or phone bill. Section 1405 of the bill permits the toll revenues to also be used for transit within the Interstate corridor. And if the state DOT certifies annually that the rebuilt Interstate is being adequately maintained, surplus toll revenues could be used for any highway (Title 29) or transit (Title 43) project statewide.
By contrast, the Reason Foundation’s Value Added Tolling principles that I wrote about last month would limit the toll rates to the amount necessary to finance construction and/or reconstruction costs, proper maintenance, and future improvements to the Interstate corridor or set of Interstates. That would keep the toll rates lower than in the states that use toll roads as cash cows to fund other transportation and non-transportation projects. Those lower toll rates would reduce the extent of diversion of traffic from tolled Interstates to parallel non-tolled highways. In other words, eliminating revenue diversion would reduce traffic diversion.
Not surprisingly, the trucking industry’s Alliance for Toll-Free Interstates has already attacked the Administration’s tolling proposal, trotting out obsolete arguments based on 20th-century cash tolling, traffic diversion, etc., as well as calling tolling an “inefficient” means of highway funding compared with unsustainable fuel taxes. It would be far more constructive if our trucking friends put aside red herrings and focused on their shared interest with other highway users and state DOTs: coming up with a viable, fair, and user-friendly approach to reconstructing and modernizing the Interstates. We have offered Value Added Tolling principles as a way to take seriously the expressed concerns of highway user groups, including the trucking industry. I’m looking forward to serious dialog with leading companies and organizations in this industry.
Colorado Dust-Up Over State’s First Toll Concession
After a 10-year effort that included a detailed environmental impact statement for the “U.S. 36 Express Lanes Project,” Colorado DOT’s toll authority, High Performance Transportation Enterprise (HPTE), approved a 50-year toll concession for the project in late February. Under the deal, a competitively selected consortium headed by Plenary Group will finance, build, operate, and maintain express toll lanes on U.S. 36 (the former Denver-Boulder Toll Road from the 1950s), and will take over responsibility for tolling and maintenance of the existing HOT lanes on I-25 between downtown Denver and U.S. 36.
But final approval of the concession agreement by HPTE and the subsequent March financial closing were threatened by last-minute protests led by local environmental group Drive Sunshine Institute. Its nominal objection was that the deal did not allow alternative-fuel vehicles to use the new lanes at no charge, but it happily worked with right-wing populists to demonize the project as a “foreign takeover of our highways” (Plenary headquarters is in Australia) and to attack HPTE for having concocted a “secret deal” for this purpose.
Although the deal went through, the vociferous opposition stirred up by the protestors has prompted bills now being debated in the legislature that would put new restrictions on future toll concession projects. Several of these are innocuous, such as requiring a public hearing when such a project is being considered, when a draft RFP has been prepared, and before the final RFP is issued. After financial close, it would also require posting the terms of the concession agreement online, a basic transparency practice that most states already follow.
But in two other areas, what the bills propose would reduce the attractiveness of doing concession deals in Colorado and reduce the performance of express toll lanes (ETLs) in ways that would undercut their financial feasibility and customer appeal. Since a number of other ETL projects are being planned in metro Denver, legislators need to consider the downsides of these new restrictions.
One set of constraints would prohibit concession agreements containing any of the following, unless the legislature approves exceptions via a joint resolution:
- A term longer than 35 years;
- A non-compete clause that either prevents CDOT from building a competing project or calls for compensation to the concession company for lost revenue from such a project;
- A requirement that HPTE compensate the company for lost revenue due to lane closures for emergencies or severe weather.
For a recent conference presentation, I reviewed the past 20 years’ of US experience with highway concession projects, to develop lessons learned. For the 14 largest projects that have been financed to date, only four had 35-year terms, with the rest ranging from 40 to 99 years. Of the four with only 35, three are lower-risk because the company in those cases is being compensated via availability payments rather than taking on traffic and toll revenue risk. And the fourth went bankrupt. Lesson learned: in general, the more risk the private sector takes on (instead of having taxpayers bear those risks), the longer the concession term needs to be to make the deal work. Another lesson learned from successful deals is that they really need to be structured as partnerships, in which both parties agree on workable provisions and have incentives to make the relationship work long-term. In that frame of reference, some degree of compensation for lost revenue generally needs to be included, but tailored to the specifics of each project.
The other provisions that concern me deal with HOVs and express bus service. One set would simply require HPTE to take these matters into account in planning ETL projects. I’m sure they were triggered by HPTE’s eminently sensible decision to increase the occupancy requirement for free passage from HOV-2 to HOV-3. In most urban areas, an express lane that gives away capacity to two-person car-pools will decimate its revenue potential (making it non-toll-financeable) and greatly reduce the pricing power of its variable tolling (making it difficult to keep the lanes flowing smoothly and reliably at 50+ mph during peak periods, which undercuts both the appeal to paying customers and the potential for much-improved express bus service).
That same objection applies to one bill’s mandate that 6,000 alternative fuel vehicles be allowed to use any ETL at no charge. This could also decimate toll revenue and undercut variable tolling’s pricing power, with serious negative consequences for any project based on toll-financing new lanes. And since metro Denver has no more existing HOV lanes to convert, all future express toll lanes will have to be new construction. Preventing that from happening could well be the underlying agenda of many of the environmental opponents of the U.S. 36 concession.
Progress-and More Questions–on Autonomous Vehicles
There have been quite a few new developments in the field of autonomous (“self-driving”) vehicles since I last wrote about this subject. In addition to the recent overview report from the Eno Center for Transportation that was noted in the November issue, both RAND Corporation and Morgan Stanley have produced thoughtful papers on the subject. The RAND report is a good overview for policymakers, finding that the benefits will likely exceed the costs, but also noting the unresolved legal and liability issues that will need to be addressed. (www.rand.org/pubs/research_reports/RR443-1.html)
The Morgan Stanley Blue Paper, “Autonomous Cars: Self Driving the New Auto Industry Paradigm,” is more optimistic about near-term feasibility than I am, finding that “autonomous vehicle technology is a smaller leap for the auto industry than electric vehicles.” It suggests that much of the liability problem would be resolved if all US states adopt no-fault insurance, but that might be wishful thinking, given the mixed record of such auto insurance to date. It also suggests that autonomous vehicles will be adopted sooner in freight transportation (e.g., truck delivery) than in personal travel. And it makes the intriguing suggestion that when people can do other things while they drive, more of them will live further from the denser parts of urban areas-“suburbanization as a future megatrend.”
I remain skeptical of the vision of “a world where nobody owns their own car,” to quote from the headline on a February Atlantic Cities piece by Eric Jaffe. But at least some researchers are now attempting to simulate such a future. In particular, if everyone relies on “robotaxis” for their commuting trips, what happens at rush hour when practically everybody summons one for their journey home? Chunka Mui reported at Forbes.com on recent simulations by a team of MIT and Stanford researchers. They used actual travel patterns and road networks of Singapore and New York City. In the Singapore simulation, they found that a fleet of 250,000 robotaxis could replace all current modes of personal transportation (cars, taxis, and transit, mostly), with a maximum wait time of 30 minutes during rush hours and with door-to-door trip time comparable to today’s. Well, excuse me, but if the typical one-way commute in Singapore is 30 minutes, that would mean doubling the overall travel time during peak periods, compared with zero-wait for your own car today. Not exactly a crowd-pleasing scenario.
In a less-ambitious exercise, the same team’s New York simulation sought only to replace the current taxi system with robotaxis. For this more-limited assignment, the results were a 40% reduction in fleet size, due to less cruising to look for fares. It would have been interesting to see the impact of replacing the current NYC transit system with robotaxis, along the lines of the Singapore simulation.
UT Austin Prof. Kara Kockelman and grad student Daniel Fagnant simulated weekday travel in a 10-mile by 10-mile portion of Austin, TX. Their conservative model looked at the impact of 5% of daily trips being made in shared autonomous vehicles. Their model found that each shared AV replaced 11 personally owned vehicles, with AV users having to wait an average of only 20 seconds before their robotaxi showed up. Of course that is hardly the same as the situation of everybody relying on robotaxis during rush hour. Interestingly, their simulations found that total vehicle miles of travel increased in the 5% scenario, due to all the driving the unoccupied AVs made to reposition themselves for their next trip. In fact, Kockelman told Eric Jaffe, “Instead of [each AV] driving 100,000 miles over 10 years [like today’s personally owned cars], they’re driving 100,000 miles every year.”
One other interesting but ominous point emerged in several of the articles reporting on recent AV research. Quite a few researchers make the assumption (or recommendation) that because AVs on average will be much safer than personally driven vehicles, non-AVs would be banned once AV market penetration is high. That strikes me as a typical ivory tower assumption, akin to the notion that has proved so harmful to progress in the mileage-based-user fee debate: that the only way to implement MBUFs is with a government-mandated GPS box in each and every vehicle. It would be interesting to see some large-scale survey research on how average Americans would view a future that banned personally driven cars.
Express Bus and Express Toll Lanes for Tappan Zee Bridge
Construction of the $3.9 billion replacement for the aging Tappan Zee Bridge across the Hudson River is proceeding, under a design-build contract. Nevertheless, two large unknowns hang over the project: how it will be paid for, and what kind of transit accommodations it will provide. Both are political problems, but there’s a single solution for both.
The bridge is owned by the New York State Thruway Authority, and was originally financed via toll revenue bonds. The Authority’s Executive Director has assured legislators that “The intent is to pay for the entire cost of the [replacement bridge] by [toll] increases at the bridge itself.” The current toll is just $5 round-trip, compared with $13 at the George Washington Bridge, the next crossing to the south. The political rationale for keeping Tappan Zee’s toll so low was that commuters using the GW Bridge have good transit alternatives, unlike those using Tappan Zee.
And it was demands for a rail transit line to be included in the new bridge that nearly killed the project several years ago, when cost estimates for the bridge plus a rail transit line were in the $10 billion vicinity. The final compromise leading to the construction going forward was the Thruway agreeing that during peak periods one lane of the new bridge would be reserved for express buses.
So far, the only financing that has been secured is a $1.6 billion TIFIA loan; the initial construction costs are being covered by 5-year bonds that the TIFIA loan will replace. No overall financing plan is yet in place, because that would require disclosing what the new toll rates would have to be. A March 25, 2014 New York Times article noted that the $3.9 billion cost figure does not include interest costs over the likely 40-year bond period. The reporter included a back-of-the-envelope estimate that amortizing $3.9 billion would require a round-trip toll rate of $11. Announcing such a figure in an election year would likely cause a political furor.
So here is my unsolicited advice to the Thruway Authority. First, express bus service on the bridge is a good idea, but taking a whole lane out of service during rush hour for this is very wasteful, precisely at the time of day when demand to use the bridge is highest. Instead of an exclusive bus lane, make it a virtually exclusive bus lane-i.e., an express toll lane with variable pricing. That would harness the pricing power of variable tolling to ensure that the express buses and paying customers would have faster and more reliable trips across the bridge than might otherwise be possible. And whenever the express-lane tolls were higher than the basic toll for the regular lanes, they would be adding to the revenue stream needed to pay for the capital and operating costs of the bridge.
Oneother suggestion to ease the sticker shock of the new-bridge tolls: develop a 40-year toll schedule that is inflation-adjusted. That way the early year tolls could be lower, in exchange for tolls gradually increasing over time, but no faster than the consumer price index. In other words, in real (inflation-adjusted) terms, the tolls would be flat.
This is hardly rocket science. Express buses in express toll lanes are in use in a number of metro areas, taking advantage of the synergy between priced lanes and express buses. And CPI-adjusted tolls are being phased in on a number of U.S. toll roads, especially those being developed under toll concession agreements. New York State has taken a modest first step into 21st century highways by finally making use of design-build contracting for a major project. There’s no reason why the new Tappan Zee couldn’t introduce several other innovations to New York State.
Tolling Interoperability via Mobile Phones
Efforts are under way both nationally and regionally to bring about interoperability among all-electronic tolling systems. This involves not only upgrading toll roads’ systems to read transponders from elsewhere but also to interconnect back-office operations so that a customer needs only one toll account and gets only one bill, regardless of where he or she drives. Texas and Oklahoma are close to making their systems interoperable, and the Kansas Turnpike in March announced plans to be interoperable with them by the end of this year. Florida and North Carolina have done likewise, with Georgia expected to join them soon, and all three plan to become interoperable with the 15-state E-ZPass group in the Northeast and Midwest. At the national level, the Alliance for Toll Interoperability is developing the first toll transactions matching hub as a pilot project. And the International Bridge, Tunnel & Turnpike Association has an Interoperability Steering Committee working to define an open protocol for a national transponder (toll tag).
A different approach is being pursued by a Florida-based startup company called GeoToll. Their approach is to use smartphones as a way to give individual customers an interoperable solution, without the need for toll operators to reconfigure or replace much of the tolling equipment in place at hundreds of different toll roads and bridges across the country. GeoToll would provide a sticker tag (similar to what many electronic toll systems already use) to be affixed to the customer’s smartphone, plus a tolling app the customer would download. When the car approaches a tolling point, the phone’s GPS knows its location, which tells GeoToll’s software which toll system the vehicle is approaching, and configures its communications with the tolling system accordingly. As the company puts it, “We bring the protocol to the lane; no need to change road-side equipment.”
I spent about an hour at GeoToll earlier this year, getting an overview of the system and watching a brief demonstration of how it works. I missed their two-hour webinar last fall, but you can read the account of it provided by Tollroadsnews.com (Sept. 18, 2013). CEO Tim McGuckin told participants that GeoToll will guarantee payment (to the toll road operator) of all tolls incurred by its account-holders, as well as taking responsibility for collecting or absorbing the costs of bad accounts. He expects GeoToll’s presence in the marketplace will increase the fraction of toll road customers paying via transponder, as opposed to the more expensive license-plate imaging. (www.tollroadsnews.com/node/6742)
GeoToll has done field tests with two toll operators thus far, Washington State DOT and another which for now remains anonymous. McGuckin told Tollroadsnews.com in a February interview that the WSDOT tests went very well, with “excellent” results, reading at the same level of accuracy as standard transponders issued by WSDOT. I saw some of those test results in my visit with the company in March. Further deployment tests are planned for this year with three other toll agencies in different parts of the country.
Skeptic that I am, I asked an independent electronic tolling expert for his thoughts, and he raised a number of cautions. I reviewed these with McGuckin and thought he provided reasonable answers-though obviously this will be determined in the marketplace. I have space for just two of these issues. First was the concern that it will take a lot of time and money to fine-tune the system for each separate toll operator. McGuckin points to a national toll industry certification program called OmniAir Certification Services as the starting point. Once GeoToll has that, each agency will indeed require acceptance testing, but there are only a handful of different tolling protocols around the country, with E-ZPass by far the largest. Illinois Tollway has volunteered to be the pilot site for the E-ZPass Group, smoothing its entry to other members. WSDOT is playing the same role for the 6C group of toll operators.
Another question was what happens if there is more than one smartphone in the vehicle as it approaches a tolling point. McGuckin notes that there is only a potential concern if more than one is equipped for GeoToll, but even then the tolling system would read them both but give them identical time stamps, which GeoToll would process as a single transaction.
My assessment at this point is that GeoToll looks promising as a meaningful step towards national electronic tolling interoperability. McGuckin suggests that with 3,300 toll lanes in the United States today, it would be very costly to replace the equipment in all of them with new multi-protocol readers. The expense of those readers would be less if all the operators agreed to only two or three protocols, but still a big expense for most. Some of the largest operators may opt for multi-protocol readers, but many others might decide to promote customer use of GeoToll as an alternative-assuming it works as well as the company expects.
Note: I don’t have space to list all transportation conferences that might be of interest. Below are those that I or a Reason Foundation colleague are taking part in.
TRB Fifth International Finance Conference, July 9-11, Beckman Conference Center, Irvine, CA (Adrian Moore and Robert Poole speaking). Details at www.trb.org/conferences/finance2014.aspx.
Background on Interstate Tolling. Washington Post transportation reporter Ashley Halsey III provides an excellent overview of the law and history of tolling on Interstate highways, as well as the specifics of the Administration’s Interstate tolling proposal in “5 Things to Know About the Obama Administration’s Highway Tolling Proposal.” It ran in the Post on May 2nd. The URL is too long to include here, so just Google it.
New Version of Delaney Infrastructure Finance Bill. Sens. John Warner (D, VA), Roy Blunt (R, MO), and Michael Bennett (D, CO) are among the cosponsors of a Senate version of the infrastructure finance bill introduced in the House last year by Rep. John Delaney (D, MD). Their Partnership to Build America Act includes the key features of Delaney’s HR 2084 that would create a $50 billion nonprofit infrastructure fund, similar in structure to the Ex-Im Bank. The new Senate version would expand the scope to include providing loans directly to P3 developers, whereas Delaney’s original would have assisted only municipal bond issuers.
Express Toll Lanes Advancing in Florida. The five-year project to reconstruct and modernize I-595 in the Fort Lauderdale area wrapped up on schedule in March. The revamped expressway now includes three reversible express toll lanes in the middle. The $1.8 billion project was developed and will be operated by ACS Infrastructure under a 35-year concession. And in late April, Florida DOT announced the winning bidder for a $2.3 billion concession project to reconstruct 21 miles of I-4 through Orlando, adding two express toll lanes each way. Winning the bid was the team headed by Skanska Infrastructure Development.
Seven Reasons for Costly Infrastructure Projects. Scott Beyer pens a cogent piece documenting seven regulatory factors that significantly increase the cost of federally funded transportation projects, including Buy America and Davis-Bacon laws. Beyer concludes by noting that “Even those who support [achieving various social goals] must acknowledge that funding them through transportation policy can come at the expense of actually building or properly maintaining infrastructure.” (www.theatlanticcities.com/jobs-and-economy/2014/04/7-reasons-us-infrastructure-projects-cost-way-more-they-should/8799)
Good Reading on Energy Policy. Matt Ridley, one of today’s best thinkers on science policy, had an excellent piece on energy and sustainability in the Saturday Wall Street Journal on April 26th. “The Scarcity Fallacy” brings an economic perspective to the question of reliance on various energy sources, whether for transportation or other uses. It challenges the conventional wisdom, but I think Ridley is right. Well worth reading.
Los Angeles Express Lanes Made Permanent. The one-year experimental express toll lanes on I-10 (San Bernardino Freeway) and I-110 (Harbor Freeway) were unanimously approved as permanent by the board of the L.A. Metropolitan Transportation Authority on April 24th. This action must still be blessed by the state legislature, which had mandated the one-year status in approving the project. After a difficult start-up, the lanes (converted from HOV lanes) have proved very popular. Metro had expected to sell 100,000 FasTrak transponders, but has sold more than 260,000. Express lane users are saving meaningful time in their commutes, and both vanpool and express bus service have increased significantly, thanks to the shorter and more reliable travel times in the priced lanes.
An Alternative View on Twin Cities 2040 Plan. Katherine Kersten of the Center of the American Experiment published a long and thoughtful commentary on the Met Council’s new 30-year comprehensive plan for the Twin Cities, which is based on Smart Growth concepts. It’s far too long to summarize here, but I’d venture to say that a similar critique could be made of similar plans in other large metro areas. I commend the Minneapolis Star-Tribune for giving such generous space to this well-argued contrary view. (www.startribune.com/opinion/commentaries/254975911.html)
Cities as Labor Markets. Alain Bertaud is a scholar at NYU’s Marron Institute whose work on urban transportation policy I have cited in the past. His new working paper, “Cities as Labor Markets,” is one of the best think pieces I’ve read in years on the relationship between an urban area’s economic vitality and its transportation system. Bertaud is a former World Bank urban planner who is working on a book on this subject. Go to http://marroninstitute.nyu.edu/news-press/alain-bertaud-cities-labor-markets.
Queensland Motorways Privatized. The Australian state of Queensland has auctioned off concessions ranging from 37 to 50 years for a set of Brisbane toll roads known as Queensland Motorways. The winner was a consortium led by Transurban and also including a major pension fund. The 70 km system includes the Clem 7 Tunnel, the Gateway Motorway, and several toll bridges. The winning bid was $6.6 billion.
Two More Toll Systems Going All-Electronic. The Sawgrass Expressway in Broward County (Ft. Lauderdale), Florida and the toll roads in Orange County, California are two of the latest systems to eliminate cash tolling by fully embracing all-electronic tolling. Sawgrass made the transition on April 19th; the Orange County transition will be completed in May.
Commuting in America 2013. The 16 brief volumes comprising the latest in the Commuting in America series (by Alan Pisarski and Steven Polzin) is available online from the American Association of State Highway & Transportation Officials. There is no better source for definitive data and interpretations of American commuting patterns. Go to: http://traveltrends.transportation.org.
“Given the situation at the federal level with the uncertainty facing the Highway Trust Fund . . . we do believe that part of our responsibility is to help states and local project sponsors develop new options, new sources of revenue. . . . We would never tell a state or local project sponsor to toll, but that option is increasingly becoming something that states are interested in, and we’ll consider finding ways to help when that’s an option states want to consider.” And, “We certainly want the private markets thinking about ways to plow their assets into American infrastructure.”
-Secretary of Transportation Anthony Foxx, interview with Yonah Freemark, April 25, 2014
“Over the course of a 75- or 99-year concession period, the [Chicago] Skyway and the Indiana Toll Road will certainly need to be rebuilt at least once and perhaps twice. The cost of this ultimately will be paid by the future users of these facilities, but this is both fair and efficient from the standpoint of intergenerational equity. In addition, the risk that the facilities will actually be generating revenue into the 22nd century will be borne by the concessionaire, not the government. If, in the distant future, people are no longer driving private vehicles on dedicated strips of asphalt, financial viability will be more the concern of private investors than of public institutions.”
-Richard Little, former director, USC Keston Institute, “Why the US Needs More Skyway Deals,” Infrastructure Investor, March 2014
“Maybe there’s a simple explanation for the auto industry’s widely reported worry that today’s teenagers don’t care about driving, and 20- and 30-somethings don’t seem as gaga over cars. Maybe Millenials can’t afford a car-at least not the new car of their dreams-according to a couple of recent pieces of research. ‘The myth is that younger buyers do not purchase new vehicles because they no longer care about them. It could not be further from the truth,’ said a study by Strategic Vision, Inc. ‘In reality, they tend to feel a deep emotional connection to their vehicles and would love to buy a new one. The main reason they do not is because they cannot afford to do so,’ the research firm said.”
-Jim Henry, “Sticker Shock Could Be Why Gen Y Is Supposedly Out of Love with Cars,” Forbes.com, March 31, 2014
“If autonomous driving is to change transportation dramatically, it needs to be both widespread and flawless. Turning such a complex technology into a commercial product is unlikely to be simple. It could take decades for the technology to come down in cost, and it might take even longer for it to work safely enough that we trust fully automated vehicles to drive us around.”
-Will Knight, “Driverless Cars Are Further Away Than You Think,” MIT Technology Review, Vol. 116, No. 6, August 2013