National Economic Education Delegation

<subject matter> NarrativeDate: May 26, 2023

Instructions to Presenter:

1) This slide deck is intended to be used as:a. A resource for ease of presentation preparation

i. Feel free to use the slides as they are currently assembled, to reduce their number, or to substitute your own w/in the NEED template.

ii. It is vital that the content of the slides not be altered so as to change the nature of the information presented.

1. If you disagree with the content, please feel free to discuss with Jon (NEED’s Executive Director jon@needelegation.org). It is entirely possible that a revision is in order.

2. Please do not extend the presentation to content that is not currently included in the slide deck w/o discussing with NEED’s ED. Alternatively, the content can be verbally briefed with the disclaimer mentioned below.

b. An indication of NEED’s statement on the policy issuei. NEED’s statement is not to be put across as an endorsement of any

particular policy, but merely what the economics profession thinks it knows about the policy issue.

2) Statement disclaimera. To the extent that you express views that are not contained in this presentation,

it is very important that you inform the audience the view expressed is your own and not that of NEED.

i. We would like to discourage you from going beyond the slide deck, but understand that it will sometimes be impossible to avoid.

ii. Please try very hard to not be terribly controversial.b. NEED’s slide decks are intended to take the audience to the brink of policy

selection. They are intended to present a common base of understanding in the economics profession.

c. Ethics, morality, and values take you from economics to policy prescription.i. If you make this leap, please be very clear that the view is your own.

Slides:

1. Opening slide

<brief summary and opening>

2. DO NOT DELETE: National Economic Education Delegationa. Brief discussion of what NEED is and NEED doesb. Use your judgement for what should be said.

3. Who we are?a. 45 honorary board – 3 Nobel prize winners, 6 former chairs of council, and 2

former Chairs of the Federal Reserve.b. 367 delegates, one in each state.c. 42 Global Partners

4. Where are we?

5. DO NOT DELETE: Credits and Disclaimer

6. Outline

There is a great deal to talk about with regard to autonomous vehicles. Today, we’re going to talk a little bit about the different worlds that we might experience with the adoption of autonomous vehicles and how we might get there.

Those possibilities are likely to be policy driven, so we will talk about some planning and policy issues as well as major economic changes that we might experiences – things like massive employment disruptions, changes in government finances and disposable income.

At the end, we’ll discuss aspects of the environmental implications of autonomous vehicles.

7. Autonomy Levels

It’s worth taking a minute to talk about different levels of autonomy. There isn’t just a choice between what we have now and full autonomy.

There are currently efforts underway both to gradually transition to fully autonomous vehicles and to introduce elements of autonomy – adaptive cruise control – lane management – automatic breaking.

There are a pair of classifications that are largely the same, ranging from fully human controlled vehicles, level 0 in both classifications, to full autonomy, level 4 in one classification and level 5 in the other.

The levels between include the elements mentioned a minute ago, which bring vehicles to level 1. Adding more features that allow the automobile to operate, but generally with a human behind the wheel, if not touching it, comprise levels 2, 3, and 4, which all involve an element of human input and attention, if only during inclement weather in the case of level 4.

8. Growth Path

This brings us to the transition from 0 to 4 or 5. There are a variety of estimates of the timing to full autonomy being prevalent on highways and roads. In particular, McKinsey has an estimate of near ubiquitous autonomy in the year 2050.

While it is possible that this is the case for the country as a whole, it is worth pointing out a couple of things. First, autonomy is going to happen much more quickly in some places than others. In particular, congested high cost urban areas are ripe for conversion, while rural areas are likely to take much longer.

It is also worth noting that companies like McKinsey have a track record of missing the rapid adoption of transformative technologies, which autonomous vehicles most assuredly are.

9. McKinsey isn’t Always Spot On

Not to impugn the organization because there are innumerable reasons for the large consultancies to be conservative in making such estimates, but they have missed the mark dramatically in the past.

In particular, in 1980, AT&T hired them to forecast cell phone penetration in 2000. Turns out they missed the mark by a factor of more than 100! They forecast 900,000 when the reality was 109 million.

It’s not clear that the same forces will drive rapid adoption of autonomous vehicles, but for reasons that will be outlined in this talk, it is entirely possible.

10. Two Important Questions

With regard to autonomous vehicles, there are a variety of important questions, but chief among those are when will TaaS be available – as this service will be very important for encouraging adoption.

Leading us to the 2nd question of just how quick will the transition be.

11. When? What do the headlines say?

On the flipside of the McKinsey report, there have been a wide variety of headlines that proclaim the imminence of the availability of autonomous vehicles. This slide was put together

about 2 years ago and it seems that these predictions seem unlikely to come to pass and are perhaps wildly optimistic.

That said, the technology is very nearly here. Waymo, Google’s self driving spinoff has been testing such vehicles very successfully on public roads for some time now. There have been accidents, but each has been of very little consequence.

Tens of billions of dollars are being invested by a wide variety and large number of companies all over the world.

12. Wildly Optimistic, But 40+ corporations are working on Autonomous Vehicles

Most of the world’s largest corporations and nearly all of the world’s major automotive companies are currently working on autonomous vehicles. Here is a smattering of titles.

<click – goes to next slide, which clicks through quite a few images>

13. Wildly Optimistic, But 40+ corporations are working on Autonomous Vehicles

<just click through…..>

14. When? What is Possible?

Given the rate of progress toward autonomy, it is not implausible that autonomous vehicles could be available by 2025.

Given previous research on transformative technologies, it is not implausible that penetration could reach 95% by 2035. With the last 5% being very difficult to achieve. The last 5% belongs to people in professions with special needs, such as hauling around tools for professional use, or in rural areas, where TaaS might not be a terribly profitable enterprise.

This forecast seems quite optimistic and dramatic, but history suggests it is possible. If we look back to the introduction of the automobile in the first place, there was very rapid adoption in major urban areas. Downtown New York City went from nearly all horses to nearly all cars in the matter of just 10 years.

Between 1900 and 1920, the number of automobiles in the United States went from 8,000 to 8 million.1 That is not the pace Something seems to be missing here

Many look at the relatively slow adoption of electric vehicles and use that as an indication that AVs will be slow to get traction. As we will discuss shortly, adopting AVs actually involves saving money, as opposed to EVs, which involve spending more on a new car than just buying an

1 Samuel I. Schwartz, “No On at the Wheel: Driverless Cars and the Road to the Future”, 2018, pg. 23.

internal combustion engine. Adopting AVs saves money while EVs cost money, at least up front, if not over the life of the car.

15. Forecast

In 2017, the group RethinkX put out a forecast for AV adoption. This forecast reflects their view that there will be a very rapid adoption of autonomous vehicles. In particular, they indicate that within 10 years of the introduction autonomous vehicles, they will account for 95% of total vehicle miles traveled.

The orange in the graph reflects miles contributed by autonomous vehicles. Beyond the time frame of their widespread use, it is also a feature of their forecast that total vehicle miles traveled will increase relatively rapidly.

We’ll come back to this issue later in the presentation, but fundamental to economics is when the price of something goes down, more of it is consumed. So it will likely be with autonomous vehicles – both that the cost of traveling a mile will fall and that the number of miles traveled will go up.

16. Step 1: Ride Sharing Tests

Autonomous vehicle developers are starting to test drive their vehicles on city streets with ride sharing opportunities. Waymo (Google) has a program in place with Lyft in Phoenix.2 Hyundai is planning a free robot service in Irvine, CA.3 So foreign companies are actively starting service, both in the US and abroad. In Singapore, NuTonomy (spinoff from MIT) has lined up with Grab, the Asian Uber – as of 2016.4 Grab intends to launch full autonomous service by 2022.5

Many other companies are also starting to provide these services on a test basis.

17. What will the future look like?

A big question about autonomous vehicles mean for the future. What exactly is the future going to look like?

2 https://techcrunch.com/2019/05/07/waymo-and-lyft-partner-to-scale-self-driving-robotaxi-service-in-phoenix/ Lyft has a similar arrangement with Aptiv in Las Vegas, which as of April, 2019 has provided 40,000 paid autonomous rides.3 https://www.theverge.com/2019/10/25/20932237/hyundai-self-driving-car-ride-hail-irvine-date#:~:targetText=Hyundai%20will%20launch%20a%20free,over%20the%20next%20five%20years.4 https://www.theverge.com/2016/9/22/13019688/singapore-self-driving-car-nutonomy-grab-ride-hail-test5 https://southeastasiaglobe.com/grab-aims-for-self-driving-taxis-in-southeast-asia-before-2022/

18. This:

From a visual perspective, it’s going to look a lot like these images. Cars will be able to “see”. This will be facilitated with other equipment attached to the car – though how visible this equipment will be isn’t clear. For instance with Tesla, the equipment is largely invisible.

The interior of a car will be very different, with no steering wheel, no pedals, no driver’s seat. The functionality of the interior of the car will be dramatically different.

19. But, will it be heaven or hell?

From a more practical standpoint, it is important to contemplate the implications of autonomy for mobility, for climate change, for accessibility of transportation. Autonomous vehicles obviously have the potential to improve all of these things, but this won’t necessarily be the case.

It is worth considering two polar opposite possible worlds with autonomous vehicles: heaven or hell. As with so many things, reality will likely be somewhere in between, but it is worth contemplating how good and how bad things could turn out to be. With these outcomes in mind, policymakers can intervene to get us as close as possible to heaven.6

20. Hell

One end of the spectrum is a driving environment that isn’t entirely different from what we have today. In particular, AVs will be individually owned and they will largely be powered by an internal combustion engine (ICE).

What makes this hell are the following implications: increased VMT, with increased pollution. Potentially increased congestion, and a continuing need for parking.

The potential need for parking is obvious. There will still be an enormous number of vehicles in service and they will have to be somewhere between trips for their owners. This parking could be off in the boondocks, but again, we are considering the worst case scenario, so people could want them parked nearby. In this case, parking will continue to take up valuable real estate (both on the streets and in terms of structures) in urban and suburban areas.

While autonomous vehicles have the potential to reduce congestion – a lot of congestion is a result of human beings being bad drivers – the increase in VMT could well increase overall congestion. With this increase in overall congestion, you could imagine the mobility gap that could result. While those with means will ride in the comfort of their own vehicle, those

6 http://greenlining.org/wp-content/uploads/2019/01/R4_AutonomousVehiclesReportSingle_2019_2.pdf The Greenlining Institute, “Autonomous Vehicle Heaven or Hell? Creating A Transportation Revolution that Benefits All”, January 2019.

without means could be stuck in underfunded public transportation laboring through the increase in congestion.

With regard to the increase in VMT, it is worthwhile to consider a scenario that involves a single family with just three members: two working adults and one child. This is sufficient to get the point across.

21. Two Adults and a Child: Morning Miles

So, with our household with 2 adults and one child, things start at home. <click>

When somebody is ready to go to work, the next thing that happens is that the car is contacted out in its parking lot – which could be anywhere, anywhere that land is cheap. <click>

The car could be in the garage, but heck, when the car can be parked somewhere cheap and summoned, why waste all that space on a garage. More on this later.

So, the car comes to the house from parking. <click>

Suppose one of the adults is ready to go first, their destination will be an office some ways off that requires transportation to get there. <click>

The car arrives at the house and takes the adult to work. <click>

The next thing that happens is that the car returns to the home for the next person. <click> That’s trip #3 already.

Suppose it’s the child that is ready next, and she is headed off to school, which is again, not walking distance from the house. <click>

The child heads off in the car to get to school. <click> Trip #4.

The car then heads back to the house to get adult #2. <click> Trip #5.

The second adult works in some other location, again, not walking distance. <click>

The car takes adult 2 off to office 2. <click> Trip #6.

The end of the morning routine is then for the car to head back to its distant parking space.

Certainly there are ways of making this process more efficient, but who’s to say that if the market is left to its own devices that this won’t be the most cost and time efficient pattern for this family?

Consider this relative to what might well happen today. The cars are parked at home. Adult one drives just #2 to work and the car parks. The child is perhaps driven to school by adult #2, who then proceeds to their own office. This involves potentially just 3 trips by the car as opposed to 7.

And this is just the morning! <click>

From just this example, it’s clear how autonomy in hell could lead to a dramatic increase in VMT. And that many of these VMTs are what are called zero passenger miles – miles driven with no passengers.

22. Heaven

Much more fun to talk about is the best case scenario: Heaven.

In the heaven scenario, hardly anybody owns a car. Instead, most automobile transportation will be provided by private fleets, think: Uber, Google, Lyft, GM, Apple. Other than for those with specialized transportation/vehicle needs, individuals will not own their own car.

Furthermore, as we are talking about heaven, all, or the vast majority of these vehicles will be electric. They will be electric for a variety of reasons, not the least of which is the fact that an electric engine has about 19 moving parts, whereas an internal combustion engine has upwards of 1900 parts. Not the cite I have be similar

7

This also means huge changes for the insurance industry. Since it is the car manufacturer that has the most control over the vehicle, what it does in response to its environment, other than where it goes, liability will reside with the manufacturer. As a result, insurance will likely become product liability and the aftermarket insurance business will be a thing of the past.

Who knows exactly when we might arrive at this heaven world, but it could be widespread in many areas, urban areas with lots of congestion and where home prices are high, as early as 2030.

7 https://avt.inl.gov/sites/default/files/pdf/fsev/compare.pdf

23. Why is this heaven?

To better understand why this is heaven, consider that not only will the cars be autonomous, but they will be shared, connected, and green! The coming together of these technologies has the potential to dramatically change the speed and safety with which people move around.Not only will speed and safety be enhanced, but the use of the world’s resources will be dramatically reduced for mobility. Not least is the effect of there simply being fewer cars in use. Some estimates indicate the potential to reduce the stock of cars in service at any point in time by as much as 80%. This isn’t surprising given that current estimates are that individually owned vehicles are idle some 95% of the time. That is an incredibly low utilization rate.

As for green, it turns out that for an internal combustion engine, as much as 80% of the energy used to propel the vehicle is simply wasted. Furthermore, with low vehicle occupancy rates, as little as 1.5% of the fuel put into the car is actually used to propel the passengers forward.8 The rest is wasted or used to move the extremely heavy vehicle.

Autonomous vehicles will be electric in heaven. There is the potential for vehicles to be much lighter – they won’t have to be nearly as well fortified against accident as accidents will happen with dramatically less frequency. And there is also the potential to right-size the vehicles. When one person is traveling, a vehicle for one passenger will show up. When two people….and so on. This will dramatically increase the utilization of fuel that goes into the vehicle.

With regard to overall vehicle miles traveled (VMT), there is the potential for these to rise, perhaps significantly. It is fundamental to economics that when the price of something goes down, people do more of it. This will likely be the case with autonomous vehicles. More on this in a bit.

However, despite the potential increase in VMT, it is still entirely possible that congestion will be reduced. It is estimated that 25% of congestion on highways is the result of accidents. With fewer accidents, this will reduce congestion. So, in principle, with zero accidents, AVs have increased the capacity of highways by 25%. I.e., VMT could increase by 25% before increasing overall congestion.9

It is also estimated that 40% of congestion is the result of bottlenecks.10 These occur for a wide variety of reasons. Most common is when somebody touches the brakes. Most occurrences of braking on a highway are the result of human reactions to conditions. Signs with too many words on them, glare from the sun, a narrowing distance from the car in front of it. It seems likely that AVs will reduce the extent of bottlenecks and hence the resulting congestion.

8 Get reference from No One at the Wheel.9 Get reference and a reference to the freeway widening study.10 US Department of Transportaton, FHA, Office of Operations, 2012 Urban Congeston Trends Operations: The Key to Reliable Travel”, April 2013. https://ops.fhwa.dot.gov/publications/fhwahop13016/index.htm

And then there is the parking issue. As we will discuss, parking takes up an enormous amount of space in the current version of the world. All of that space could be used for much more productive purposes.

24. Transition

This leaves the question of how we get from here to there. Some suggest that the transition will be relatively quick. And that is entirely possible. We are, in fact, already in the process of transition. By that, I mean that safety features that were likely developed in the process of composing full automation are already making their way into new cars. Things like adaptive cruise control – that tracks the car in front of you – or lane control – keeping you in your lane.

So, in the short term, which we are already in, we will have more of these features available.

In the medium term, there will be a short period of personal vehicle ownership with level 3 capability. What this means is the ability to take your hands off the wheel, but while still staying alert and with hands close to the wheel.

While these capabilities are being built into vehicles available to the public, fleets of autonomous vehicles will be slowly rolling out with level 4/5 capability. There will be this transition period until full buy-in of autonomy is achieved.

At this point, the long term – and remember, the long term may be no less than 10 years away – private individuals will shed their ownership and rely instead on FAVES. Fleets of autonomous vehicles.

Often questions arise about this possibility. “But people like to drive!” My response is usually “don’t judge tomorrow’s preferences by today’s possibilities”.

People may like to drive, but they don’t like to go to the gas station, they don’t like to get their vehicle repaired, they don’t like to write the insurance check. They don’t like to wash their car, or go in for an oil change.

And they will like the freedom that comes from using a service.

Turns out many people buy their car for low frequency events. Think: I might want to go skiing for the weekend – so I’ll need a large vehicle. They then end up doing 95% of their driving in a car that is too big. With FAVES, like Alice’s Restaurant, “you can get anything you want”.

25. Economics Drives Transition: Private

This transition will be “driven” by a number of forces. These forces are both private and public.

In terms of the private forces, there is an “Adoption Dividend”. This adoption dividend means that if you give up your personal vehicle and use FAVES, you will save a bunch of money. Current estimates are that it costs, on average, $9,282 per year to own and operate a vehicle.11 This is all in, gas, depreciation, insurance, maintenance, everything. It will off course vary by the type of vehicle you have – sedan is less, truck is more, Ferrari is much more!

This figure suggests an average cost per mile of about $0.59 cents per mile. There are estimates available that suggest that the cost of using a shared AV could be as little as $0.19 per mile.12 There are quite a few things that could change this calculus – such as very little competition in the car service industry, which would cause the price to be higher, perhaps significantly higher.

There is the additional financial benefit – and practical benefit – of being able to repurpose your garage. In today’s housing market, adding a bedroom to a house raises the value of the house by about $50,000. In some places that increased value is significantly higher. For example, in San Mateo County, just south of San Francisco, it may add as much as $200,000 to the value of the house.

So, whether you turn the garage in to a new bedroom, play room, or combination of storage and something else, it is likely to increase the value of your house.

This conversion doesn’t happen now because of the belief, accurately, that the next owner of the house would like to have a garage. So merely replacing the garage with a bedroom in the absence of FAVES, will not currently increase the value of the home.

An additional part of the dividend is time recovery. This speaks to the rapidity of adoption of AVs. People will value the increased time available when riding in an AV. In the San Francisco Bay Area, more than 50% of the workforce has a commute in excess of 30 minutes. Imagine getting an additional hour added to your day!

<this is just a joke, but true>

Finally, driving your car will simply be too annoying when half the cars on the road, the AVs, are actually obeying the law!

26. Economics Drives Transition: Public

It turns out that there are massive externalities associated with human driven vehicles. Though most of you probably haven’t thought about it in these terms, but the price paid by human fallibility is quite high!

11 https://newsroom.aaa.com/auto/your-driving-costs/12 RethinkX study.

As mentioned earlier, accidents are responsible for congestion being about 25% higher than it otherwise would be. There are other costs, besides the time eaten up by congestion.

In particular in 2018 was the third consecutive year of more than 40,000 people killed in car accidents.13 Adding to this cost is the number of injuries. In 2018, there were some 2 million injuries associated with car accidents.

Furthermore, the NHTSA estimates that more than 90+ of accidents are a result of human error.14 In principle, therefore, if you take the human out of the driving equation, the frequency of accidents would be significantly reduced. This does, however, depend on the self driving technology being extremely good.

The costs associated with human driving have been estimated by Morgan Stanley. They suggest that AVs could save $1.3 trillion annually.15 Here is a breakdown of those potential savings.

27. Potential Savings

As just indicated, a huge source of potential savings is from accident avoidance. This saves about $488 billion per year. Another important category that is roughly the same size is increased productivity. This is the value of time saved by not having to be actively driving the car.

More productivity gains arise from the reduction in congestions, and there are fuel savings not only by the shift to FAVES that are electric, but also because of the reduction in congestion.

All in all, these categories add up to some $1.3 trillion in savings annually.

28. Public Policy/Planning Issues

Given the potential savings due to reduced externalities, it is essential that there be buy-in on the part of the government. It will be well worth the government’s efforts to facilitate the adoption (in the heaven sense) of autonomous vehicles.

At the same time, the transition to FAVES will result in the elimination of a large number of jobs. Given this difficulty associated with the transition, there will be some significant resistance to the widespread adoption of AVs by many government officials.

Among the many things that governments must contemplate during the transition are the following. Of the current infrastructure plans, how much of the planned development will still 13 https://www.nsc.org/in-the-newsroom/2018-marks-third-straight-year-that-motor-vehicle-deaths-are-estimated-to-have-reached-40-00014 https://one.nhtsa.gov/people/injury/research/udashortrpt/background.html15 https://www.morganstanley.com/articles/autonomous-cars-the-future-is-now

be necessary once AVs are commonplace. Second, what are the implications for public transportation? And finally, what to do with all of the parking spaces.

We will discuss each of these in more detail shortly.

29. Planning

In terms of government planning, there are many facets to the issue, but two rise to the top in terms of importance.

First, infrastructure investments need to be reevaluated in terms of time to recoup the initial investment. Many infrastructure investments are evaluated with time to a positive rate of return in excess of 25 to 40 years. Many planned investments, particularly those intended to accommodate the growth of human driven vehicles, should be reevaluated. AVs may well be commonplace long before 25 to 40 years have passed.

In other words, the planning horizon has gotten much shorter.

Second, governments should consider their planning and infrastructure development and how it will influence the transition to AVs. Accommodating the growth in human driven vehicles means that the cost of driving your own car will be lower. A lower cost of driving means that AVs will be less attractive.

Projects that facilitate private car ownership and human driving should be reevaluated in the context of the arrival of AVs.

30. Responding to the coming changes

In order to appropriately plan for the pending changes, it is imperative that transportation agencies develop a forecast for adoption. This is an incredibly complicated and uncertain endeavor, but adoption rates will likely vary significantly by location. Urban congestion areas will convert first, with rural regions following some time later.

This forecast has to be considered when considering the ROR of planned projects. Where AVs arrive sooner rather than later, the impact on ROR could be significant.

These projects include adding highway miles or expanding public transportation.

31. Encourage Change

I have already talked about the $1.3 trillion in potential savings annually, but there are more reasons to encourage adoption of AVs. They fall into primarily four buckets:

a) Mobility and Equity considerations

a. Think about the increased mobility for elderly, disabled, impoverished.b) Safety

a. the only way to reduce traffic accidents and fatalities is through government action. Could be regulation, law enforcement, or encouraging AVs.

c) Productivitya. Largely through reduced congestion and eliminating time paying attention

behind the wheel.d) Environment

a. AVs will (may) speed the transition to electric vehicles, reducing the impact of mobility on climate change.

These are all benefits to society that will either not come about, or will come about too slowly in the absence of public, or government influence.

32. Mobility and Equity

AVs have an enormous potential to enhance the mobility options for handicapped, elderly, and low income individuals. For anybody who can’t drive or who cannot afford a car, the potential is there to expand their mobility options.

In particular, for low income individuals, in the absence of a car, public transportation does not always serve their purposes very well. Public transportation tends to go between residential and commercially zoned regions. Often, low income workers need to travel from residential to residential. Taking public transportation can lead to extremely long, potentially expensive commutes for these folks. AVs have the potential to change this.

Could lead to a mobility gap.16 There is the legitimate concern that, particularly in the hell version of the world, a mobility gap can arise. Ensuring that FAVES are accessible in all neighborhoods and to people of all income categories will be a significant concern.

33. Safety and Productivity

The safety and productivity benefits of AVs and FAVES are reasonably self evident. Think fewer accidents and less time paying attention to the car while riding.

34. Environment

From an environmental perspective, there are a variety of benefits. Beyond the switch from ICE to electric engines, there are many environmental benefits. Estimates suggest that as much as 30% of the cars in congestion in cities like San Francisco are searching for parking.

16 https://www.salon.com/2019/03/05/how-self-driving-cars-could-harm-marginalized-communities_partner/

Another benefit is the freeing up of space from parking. Think about apartment buildings. Parking is often mandated to accompany new apartments. Often the space devoted to parking can exceed half the size of the actual apartment!

Or, think about churches. What does every church have right next to it? A huge parking lot, often with a bigger footprint than the church. This could be used for green space to improve the environment.

Or think about surface parking in cities. Changing this to green space has the potential to improve the environment significantly.

35. Incentives Through Policy and Planning

In order to bring about these benefits, there are a variety of potential policy options for government. These options include both carrots and sticks.

In terms of carrots, one could imagine opening up and expanding HOV lanes to be both HOV and AV lanes, where the definition of AV evolves over time as more autonomous capability is available in cars. In particular, perhaps allowing cars with adaptive cruise control into the lanes would be a good start that would encourage people to incorporate such technologies.

In terms of sticks, there are a variety, and they are perhaps best geared towards reducing individual car ownership. The various fees associated with car ownership ought to be brought more in line with the actual costs of private/individual ownership, which as we have discussed, are considerable.

36. Interim Summary

A couple of things are worth emphasizing at this point. In particular, it is likely that AVs are coming pretty quickly. How quickly, nobody really knows. That said, for something as disruptive as AVs may well be, the transition has the potential to be very, very fast.

Second takeaway is that governments, both local and otherwise, should start considering AVs in their planning now. Not only are there enormous benefits to be had from their quick adoption, but there is a lot of money to be wasted if current planning doesn’t take into consideration their arrival. Many investments currently on the books simply don’t make sense. It seem unlikely, in particular, that we will need more highway miles or wider local roads when AVs arrive. We will, however, want good roads and functioning bridges. Investments in maintenance may well be worthwhile.

37. What Changes Will This Bring?

AVs will bring about an enormous number of changes. Here, we will consider eight important categories of change. These changes range from the individual level to all levels of government, in particular with regard to transportation planning.

38. Disposable Income

The effect on disposable income stems from the hope that people will not own a car. According to AAA, which provides an annual estimate of the costs of car ownership, in 2018, it cost the average car owner more than $9 thousand per year to own a car. This cost is all in, gas, insurance, etc.

It has been estimated that the cost of replicating the average car owner’s costs of personal transportation with TaaS (Transportation as a Service) could well be less than $3,000.17

If that is indeed the case, then the average car owner will experience a net increase in their disposable income – money that was devoted to transportation, but now can be devoted to, well, pretty much anything else.

Spread across all households, this makes available more than $1 to 2 trillion in new spending in the economy – each household had an estimated 1.88 cars in 2017.18 Now, although given our habits, it seems unlikely, but is possible that we will save some of this new disposable income. Nonetheless, it will still likely be an additional $1 trillion in new spending – which will support quite a few jobs.19

39. Employment

Speaking of jobs, AVs bring with them the potential to render quite a few jobs obsolete and to reduce the demand for people in many other jobs.

Starting at the top, anybody who drives for a living will find their job at risk. This includes taxis, trucks, ambulances, delivery vehicles. Perhaps the first will be those who drive long haul trucks. Autonomous trucks may be among the first automated vehicles to really make inroads. Part of the reason for this is the amount of savings available to trucking companies if they can either omit the driver altogether or even just extend the moving hours of existing drivers.

17 RethinkX18 https://www.statista.com/statistics/551403/number-of-vehicles-per-household-in-the-united-states/#:~:targetText=On%20average%2C%20there%20are%201.88,disposal%20in%20that%20same%20year.19 See Tony Seba and James Arbib for a comparable calculation from RethinkX.

Next will be jobs associated with the manufacture of cars. By some estimates, the number of cars in service at any given time may decline by 80%.20 At the same time, their lifespan is likely to increase.21 Both of these forces indicate fewer cars being produced in a given year.

Beyond the production of vehicles are parts suppliers. There are again two significant forces that will lead to a reduction of jobs in this sector. First, fewer cars means fewer parts. Second, a transition to electric vehicles means fewer moving parts per vehicle.

Next up are jobs associated with the functioning of automobiles. Electric vehicles will put service stations at risk. Fewer accidents will dramatically reduce the number of vehicle repair shops and jobs associated with them. What jobs in vehicle repair and maintenance there will be will likely be at the providers of TaaS, rather than independent companies, such as often the case currently.

Fewer accidents, less speeding, the elimination of DUIs all imply less demand for police and fire services, so police and fire jobs will be at risk – not of obsolescence, but of reduced demand.

Another level is employment in hospital emergency rooms. With as many as 2 million fewer injuries from car crashes, emergency rooms will be less busy than they are today.

With many fewer injuries, this reduces the demand for medical supplies, so jobs in this sector may well be at risk. Although this effect may well be small and a slowing of growth rather than an observable reduction.

These are some of the obvious jobs that are at risk from widespread use of AVs. It is likely that there are others that are more difficult to detect.

With regard to dislocation and finding a new job, most of the drivers who are at risk ar making an average of $13. If laid off, not only are they spread throughout the economy, so no particular region is specifically at risk, but they will take less of a hit than manufacturing workers should the be laid off in taking a $10 an hour job at Walmart.22

<get study from Bill Fox on estimates>

40. Employment (con’t)

As with most economic change, the jobs that are at risk are much more obvious than the jobs that will be created. Fundamentally, AVs are a significantly labor saving technology. In fact, several of those pursing the technology are doing so with the intention of replacing human drivers in their current service with automation.

20 Reference from No One at The Wheel21 RethinkX?22 Lawrence Katz in “Driverless Future?”, https://prospect.org/economy/driverless-future/

Also as with most economic change, there will be jobs created that are extremely difficult to anticipate. Examples of jobs that will surface are IT jobs. As a technology that will be widespread throughout the country, there will be a need for individuals to work to maintain the vehicles. These will be significant in # and will primarily involve relatively skilled individuals.

Think about the adoption dividend that was mentioned just a few slides ago. If households have in excess of $6,000 in additional income, they will surely spend a significant chunk of that. These purchases will be of goods and services in all sectors of the economy.

So, the answer may well be that jobs lost to AVs will be replaced by jobs throughout every other sector of the economy.

The trouble with this replacement is that many of the workers losing their jobs may not have the skills necessary to find employment in an expanding sector. As a consequence, a significant amount of worker training will likely be necessary.

Much of this retraining may have to be provided by government organizations. With the rapid pace at which autonomous vehicles appear to be coming, it is probably time to start thinking about what these programs might look like now.

41. Government Finances

With autonomous vehicles, many sources of local revenues will be at risk. With cars that don’t speed or otherwise drive poorly, there will be fewer tickets for moving violations. With cars that don’t drink, there will be fewer DUIs. With potentially dramatically reduced parking, parking revenues will decline significantly.

Revenues from all of the ancillary services that are currently performed – fuel, repair, maintenance – will also decline.

At the same time, the space vacated by these shops may be filled by some other business. A significant problem may well be environmental cleanup, however.

With less parking, there will be more space for commercial, retail, or residential space. Consider, for example the massive parking lots around shopping malls. It would seem reasonable to expect that some of this space would be turned into residential or additional commercial.

Not only will revenues be significantly altered, but expenditures will as well. As mentioned, perhaps significantly less labor will be required (police and fire, in particular). But perhaps a larger effect could be related to road development and maintenance.

AVs will lead to the more efficient use of existing roads. Lanes can be narrowed, if need be to perhaps provide more lanes, and throughput of existing lanes will also be improved – for all of the reasons already expressed.

This will reorient funds that are currently earmarked for expansion to maintenance. There is also likely to be significant lobbying from TaaS providers for maintenance. Cars require less maintenance with roads that are better maintained. Potholes remain an enormous public cost.

At the same time, the TaaS providers may well provide an obvious source of revenues. Taxing them for the publicly provided services that they use will likely be a much less controversial undertaking than taxing individuals. These providers may also welcome the taxation as a way of encouraging more efficient use of their vehicles, so it is ultimately a profit maximizing strategy.

As with the rest of this conversation, there are a great many unknowns here. It is well worth contemplating them in advance of the mass arrival of AVs.

42. Transportation

It is axiomatic to economics that when the price of something declines, more of it is likely to be purchased. So it is with transportation services. It is quite likely that the cost of moving a mile in a vehicle will be lower with TaaS than it currently is.23 Accordingly, people will travel more miles. In particular, living further from your place of employment will become less costly, so daily commutes will increase for some or many.

There will also be some amount of zero passenger miles that are generated. To some extent these will be offset by a reduced search for parking, but will likely exceed that amount, so will add to vehicles moving on the road.

At the same time, as mentioned, there will be a declining demand for additional road miles, other than to serve currently unserved areas. In particular, highway widening will be much less attractive. The more efficient flow of AVs will significantly increase the efficiency with which cars use current road. This effect is true both front to back – cars can travel closer to each other in the same lane – and side to side – cars can be closer to each other when traveling in different lanes.

43. Public Transportation

The effect of AVs and TaaS on the demand for public transportation is one of the truly important, if unanswered questions.

23 Reference? I reference someone in my paper but it is a secondary reference so not very good.

The appeal of public transportation is based on three factors: cost, convenience, and time. The relevance of TaaS and AVs for public transportation depends on whether or not they substitute or complement public transportation.

The answer to this is ambiguous generally. TaaS may dramatically lower the cost, increase the convenience, and reduce the time associated with an alternative to public transportation.

At the same, time, it is possible that by easily extending public transportation, AVs and TaaS could increase the convenience and time factors of public transportation. That is, TaaS may well solve the “last mile” problem, the problem of getting the individual from public transportation to their job or their home.

It seems likely that TaaS will substitute for public transportation on short trips, while it may more effectively complement TaaS for longer trips.

On a shorter trip, TaaS may well be competitive on cost, convenience and time. If there is a last mile problem on both ends of the trip, this means that public transportation is a 3-mode trip, whereas just staying in the AV will be a single mode.

44. Cautionary Tale From Long Ago

With regard to the effect of AVs or TaaS on public transportation, we can look back to the introduction of motorized transportation. In this case, trucks. I believe that there is an analogy between the earlier experience of railroads with regard to shipping and public transportation and moving people.

Back in 1894, well before the automobile was ubiquitous, but trucks were coming into increasing use, there was a significant impact on railroads as they experienced increased competition from trucking. Surely this wasn’t true on all routes, but on routes of a certain distance, where time was also competitive, led to reduced demand for rail based shipping.

It seems likely that something similar may happen with regard to public transportation.

Again, the longer public transportation routes may well experience increased demand – because of the solved last mile problem – but shorter routes, because AVs and TaaS may be cost and time competitive, could see significantly reduced demand.24

45. Infrastructure

The current focus of transportation spending is on the expansion of infrastructure, including more lane miles. Witness the very poor grades that infrastructure nationwide gets from

24 The Atlanta Regional Commission predicts, in an extreme case, that public transit trips in their region may decline by up to 42% by 2040 with the introduction of AVs. https://documents.atlantaregionalcom/taqc/217/20170-01-12/Regional_Trails.pdf

American Society of Construction Engineers. The 2017 Annual Report Card gives the nation’s infrastructure a grade of D+.25

Going forward, with AVs and TaaS, it seems unlikely that additional highway miles will be necessary in the short term. AVs are likely to lead to an improvement of throughput on existing roads – platooning and communicating, facilitates speed and safety. It’s even possible that additional lanes can be added to highways without expanding the current footprint of roads. Currently, lanes are 12 feet wide while most vehicles are 6 feet wide – allowing for significant human error that will no longer be present.

In order to facilitate the more efficient use of existing infrastructure, such spending can instead be used to make the roads smarter and towards better maintenance.

Indeed, there will likely be significant lobbying by the fleet owners (Uber, Google, Lyft, GM, …) for roads that are better maintained. Pot holes cost society significantly, though the costs are currently very widely dispersed. With TaaS, and perhaps as few as 5 owners of vehicles, the costs are no longer widely dispersed and maintenance is more likely to happen.

46. Housing

Although it might not be an obvious connection, it is reasonably clear that AVs and TaaS will have an influence on the ability to build more housing in space constrained and housing constrained regions.

In many places, it is traffic congestion that is an impediment to the construction of new housing. With autonomous vehicles, under the heaven scenario, it is entirely possible that local congestion will be reduced. It is also likely that the individual level costs associated with congestion will be reduced. If you are riding in an autonomous vehicle, you can be productive, whereas most often today, congestion is a direct impediment to productivity.

With reduced congestion, the pushback on residential construction may well be reduced, allowing for more construction.

It is also possible that because of the “adoption dividend” that we discussed earlier, the conversion of garages to bedrooms will expand the housing capacity in most housing constrained regions. As a 2 bedroom becomes a 3 bedroom, a 3 bedroom becomes a 4, the capacity to house people is increased, releasing some of the pressure to approve of new construction.

25 https://www.infrastructurereportcard.org/

47. Parking

Parking plays an enormous role in current mobility. Personal vehicles are generally idle 95% of the time. And given that parking for an individual vehicle happens in a variety of places, there is currently more than one parking place for every vehicle in use nationwide.

With TaaS, parking will become a much smaller part of modern mobility. Although demand for parking will never disappear, it will be significantly reduced. Churches will need less parking, urban centers will require less parking. Street parking will be generally in lower demand.

It is likely true that service providers may well own parking lots in strategic places. There are clearly more vehicles necessary to provide service at commute times than in the middle of the night. Those additional vehicles will likely park rather than just circling. This begs the question of where? Will there be lots in urban centers? Or will the vehicles all return to a “home base”, well away from population centers. It is too soon to answer this question.

But we can well imagine that street parking could largely be eliminated. As such, sidewalks can be widened, buildings can be larger, more greenery can be included in urban design.

Shopping malls current have an enormous amount of parking. This parking will no longer provide the service to the retailers in the mall that it currently does. This space can be used for a wide variety of other productive purposes. Perhaps for more retail, for housing, parks, or other options that I’m not creative enough to imagine.

Similarly with apartment complexes. We have already seen that most places require significant space be devoted to parking for additional complexes. This parking can be turned into additional housing, or perhaps green space.

48. Freeing up Urban Space for Parking.

The amount of space currently occupied by parking facilities is staggering in size. In Los Angeles, it is 14% of the incorporated land area. This is 200 square miles!

In San Francisco, there is enough on-street parking to parallel park a line of cars 60 miles longer than California’s coastline.

This is enough parking to cover the Presidio, Golden Gate Park, and Lake Merced.

Nationwide, estimates are that there are 500 million spaces available, but it could be as many as 2 billion. This is enough to cover the four states of Delaware, Rhode Island, Connecticut, and Vermont.

An enormous amount of space is currently allocated to parking in the United States. This is likely to change with the forthcoming changes in mobility.

49. Summary of Change

The changes coming in mobility, perhaps within the next decade, are enormous. And they bring with them enormous economic implications.

In particular, more than 10 million jobs will be put at risk. Many jobs will be created, but a mismatch between old and new jobs, both in terms of skills and geography make this an enormous public policy issue.

Local government revenues will look very different. Sources of revenues will shift away from those related to mobility – speeding and parking tickets, DUI related revenues – and expenditures will experience a similar shift. Fewer police services will be required – in particular, traffic control and monitoring.

Given the potential implications for congestion, or at least the costs of sitting in congestion, there will be positive effects on the ability of local governments to approve the construction of additional housing.

Parking conversions, freeing up space for a myriad productive uses will occur.

Similarly, because of reduced congestion and the costs of sitting in it, there will likely be less demand for expansions of transportation infrastructure, but increased demands for maintenance.

Finally, the implications for public transportation are ambiguous. It seems likely that TaaS will serve as both a competitor and a complement. Which it is likely depends on the nature of the trip. Shorter trips will find TaaS a competitor while longer trips will find TaaS as a complement, solving the “last mile” problem.

50. Environmental Implications – Depends on Heaven or Hell

From an environmental perspective, there are multiple implications. Because there will be fewer vehicles in existence, resource allocation will be improved. There will be much less steel, aluminum, rubber, and other resources in use in the automotive sector.

Travel will be more efficient. Vehicles are likely to be, at least on average, “right sized”. If fewer people are traveling, the vehicle will be smaller. Given the technology available in the car, routes will be optimized, weighing off congestion and distance, providing improved outcomes in terms of energy utilization.

As we mentioned, about 25% of the vehicles in urban congestion are looking for parking. This practice will cease, reducing VMT and providing some congestion release.

The cars will be electric, instead of ICE and the vehicles can ultimately be lighter. Without the infrastructure in each vehicle to protect its occupants from accidents – because there will be fewer accidents – the cars can be significantly lighter, improving fuel efficiency.

At the same time, the computing power available in each car will add to energy consumption both in its general operating as well as the weight that it adds back to the vehicle. It seems reasonably clear that the environmental implications will be positive, but there are both positives and negatives associated with the transition to AVs, especially if we end up in Hell instead of Heaven.

As a bottom line, it is likely to government intervention as just what state of the world we end up in.

Governments determine local zoning, design roadway alignments, invest in new capacity, design parking regulations, set transportation taxes, and invest in data and people to help manage all those decisions.

It’s a huge suite of responsibilities, and all apply in new ways with AVs. And these decisions will be instrumental in tapping all of the benefits that are available from the introduction of autonomous technology.

51. Questions?