Technology Day 2004 - "Shifting Gears”

[MUSIC PLAYING]

GARVIN: They assured me we're not going to give you a lavalier. The mics will be on. This is the Massachusetts Institute of Technology and this is how we start the day. It's my pleasure as the executive vice president of the Alumni Association to welcome all of you this morning to Tech Day 2004. The committee and staff working on the program have put together another stunning day of academic interest with time for Q&A, discussion with the presenters. And we're very excited about the program.

We are living on the edge a little bit this morning. We're scheduled at 11:45 to have our alumnus traveling the furthest for his reunion to join us from the International Space Station. And we hope that all will go well with NASA and in addition to talking about the automobile this morning to get a little report from space. There has been a change in the afternoon programming. There will be a wonderful opportunity for I think more discussion. We tend to pack these programs so tightly with speakers that sometimes speaker cancellations are a blessing.

This afternoon, we will have Norman Augustine, Dean Cayman, and Ernest Moniz, and have time to really talk about some important issues with them. I wanted to recognize all of the good friends here and the people I hope will soon be good friends for your continued support of MIT. And I wanted to especially pay tribute to one special guest.

It was noted at commencement yesterday that it was the first time in 44 years that Paul Gray was not on the commencement stage. Paul has been on that stage as a faculty member, chancellor, president, chairman, and now again as a faculty member for the last 44 years. But yesterday he was instead marching with his 50th reunion class as the class president. And I think for that remarkable achievement, we owe Paul and Priscilla a round of applause.

[APPLAUSE]

Paul and I were discussing this yesterday. I believe that must be a record, although he thinks that perhaps President Killian spent as many years on the stage. But I suspect Paul will surpass that in the future. It's my pleasure to introduce our morning moderator. The day will be split into two sessions, this morning moderated by President Vest, in the afternoon by our distinguished guest, Norman Augustine.

Chuck needs no introduction to this crowd. He is MITs 15th president. And conveniently for our topic today and our guest speaker, he chaired the President's Advisory Committee on the redesign of the space station in 1993 and '94. So he knows far more about this than the rest of us. And we're delighted to have this opportunity to feature that today. Chuck needs no introduction, but he deserves our biggest applause for his bold and disciplined leadership of MIT for the past 14 years. Without further ado, President Vest.

[APPLAUSE]

VEST: Thank you, Beth, and good morning everyone. Over the last few years you've heard a great deal about so-called disruptive technologies, phrase coined by Clay Christensen up at the Harvard Business School. Disruptive technologies are those innovations that have the potential to truly change the way we live, to transform the economy, and in the end to reshape our world.

Today's technology day program is designed to discuss what certainly has to be the greatest disruptive technology of all time, namely the automobile. Over the course of the last 100 years or so, the automobile has spawned entirely new industries, remade cities, and spurred the growth of suburbs and exurbs and come to play an increasingly important part in our daily lives.

Now I'm going to be honest with you at this point in my speech, there's a little sentence in brackets that says tell some anecdotes about your personal experiences with cars. But I figure that if I got started on our families '36 Ford, my '49 Chevy or my '49 Plymouth, my '54 Volkswagen, or even my 1985 Dodge Aries station wagon with a stick shift, we'd be here all day. So let's hear from some people that really know about this industry.

For most of this period, the United States has been thought of as a car culture. But since World War II, the private automobile has actually been transforming Europe as well. It has played an enormous role in the development of contemporary Japan and South Korea, and is now beginning to make its mark in China and other rapidly emerging economies.

There is no doubt that the automobile has been and remains a powerful symbol of personal freedom, an important engine for both economic opportunity and for leisure. But there is at the same time no doubt that it has troubling consequences as well in consumption of energy and the production of environmental pollutants and in land use and community planning. In other words, the automobile is a powerful technology that in fact is still developing and that has profound economic, but also social and policy impact as well.

If that doesn't sound like the kind of a subject that MIT ought to be concerned about, I don't know what is. In fact these are exactly the kinds of issues and problems that MIT as you know seeks to tackle, places where the world needs our expertise and our vision. So it is not surprising that we have been involved for many decades with questions raised by the automobile.

As far back as 1929, a generous gift from Alfred P. Sloan Jr. provided for the establishment of the Sloan Automotive Laboratory, which remains an international center of research in automotive and engine technologies. During the 1980s and '90s the automobile was at the center of MITs groundbreaking studies of industrial productivity and competitiveness that began to develop during Paul Gray's presidency, at a time in which as you know the manufacturing sector of our economy was deeply threatened, and today work on issues related to the automobile continues across the Institute as we develop both new automotive technologies and a new understanding of the car social and environmental impact.

The impact of this work is leveraged through a collaboration with many stakeholders from industry, from government, and from non-governmental organizations, NGOs. This morning we will hear about central challenges and opportunities before us today about the automobiles economic impact, its fundamental technologies, its impact on the developing world, and its role in individual lives. This afternoon's session will look to the future, encouraging us to think about resources, the relationship of the car to urban life, and the possibility of alternatives.

This morning our first speaker is going to be professor Dan Roos. Dan is the Associate Dean for engineering systems in the School of Engineering and the co-director of our engineering systems division. A member of the MIT class of 1961, he holds the Japan steel industry professorship of engineering. Dan is going to talk to us about the automotive sector, future challenges and opportunities. But I can't help insert Dan at this point a little anecdote that says something about the quality of your work and the work of MIT in general.

Rather early in my presidency, not too long after Dan's remarkable book, The Machine That Changed the World, about productivity and lean manufacturing in the automobile industry came out, I had a trip to Germany on behalf of MIT and had an appointment to spend an hour with the then chairman of the then company Daimler-Benz. And I had been warned up one side and down the other that everybody at Daimler-Benz recognized their company and its miserable productivity and quality ratings in the anonymous companies in Dan's book. Watch out, they said. This is a big, powerful European chairman. He's going to beat you up.

So I walked in with great fear and trepidation, said hello and so forth, sat down, and he looked over at me and he said, that book that you guys wrote at MIT, it's going to have more positive impact on this company than anything that's ever happened. And that's exactly what we want to have happen here. And I think indeed Dan's work had impact not only there, but around the world.

Following Dan we'll hear from another extraordinary MIT colleague who absolutely anybody who knows anything about engines and combustion and automotive technology knows and looks to as one of the great gurus, our colleague Professor John Heywood. John currently is the Sun Jae professor of mechanical engineering and the director of the Sloan Automotive Laboratory that I mentioned a moment ago. John will discuss future automotive technology and fuels, the options and their impact.

Following the talks, the remarks by Dan and John, the three of us will then go to the other side of the stage and we will open up for about 15 minutes of question and answer with you before the mid-morning break. Dan.

[APPLAUSE]

CREW: They're asking if you'd please take your name tag off when you speak.

[APPLAUSE]

ROOS: It's a pleasure and privilege to be here today. I'm here wearing two hats, a speaker in this technology day, and as was mentioned earlier I am also here as an alumni of MIT and a member of the class of 1961. So I thought it might be interesting to take a look at what a car looked like in 1961. The 1961 Chrysler Imperial getting about 8 miles per gallon with tail fins. Like I could have put a Cadillac in which had about three times as tall tail fins.

Now, bear in mind that it wasn't until 1965 that Ralph Nader wrote Unsafe at Any Speed. There was no Environmental Protection Agency. As a matter of fact, there was no Department of Transportation. Japanese cars were not being sold in the United States at that point in time. So things have really changed in a very, very significant way.

Chuck very eloquently mentioned why the automobile is so important to us, from a societal point of view in terms of its economic impact. From a personal point of view it provides us with mobility to get to work, to shop, entertainment, and from a societal point of view as well.

So when we think of the automobile we have to think of it in an overall context of an automotive system. There's the infrastructure upon which the vehicle is going to operate, which is being subjected to increased congestion. There's the urban environment and the impact of both the vehicles and the infrastructure on that environment. There are concerns about sustainability, the impact on our environment, particularly with regard to urban air pollution and global warming; and other societal concerns, fuel efficiency, energy conservation, automotive safety, the disposal and recycling of vehicles, and some new concerns, issues like road rage. And more recently for those of you who have been following all of the debate and concern about obesity, joining the fast food chains as one of the primary villains is the automobile, with the implication being and probably correct implementation because people are driving so much they're walking less.

So this is a technology that has had, as Chuck said, a really significant impact on all of us. And also as Chuck indicated the growth and use of automobiles is increasing all over the world. This is not a US phenomena. We have become a dependent society on the automobile. Public transportation is an important, but unfortunately it plays a minor role in terms of the amount of passenger travel that it carries. And that role is decreasing, and it's decreasing even in the major, large, high density cities of the world, the cities that were really developed around public transportation.

The public has expressed their desire for the comfort, convenience of automobile travel. They like a more low density lifestyle. And so we really have become an auto dependent society. There's no alternative to the automobile.

Today I'd like to begin looking at what's happening in the auto industry and then look at the future in terms of what auto production may look like, where the future markets will be, some of the challenges that we have with respect to highway congestion automobiles and the urban environment, and finally a few comments about the implications for MIT.

Automobile manufacturing in the US has changed drastically. In the 1980s all Japanese cars sold in the United States were manufactured in Japan. Today, the big three, Chrysler, GM, and Ford, the American producers, are closing manufacturing plants at the same time the Japanese companies are opening up manufacturing facilities in the United States, largely as a result of globalization. And it's not just manufacturing facilities. It's research and development. It's design centers so that these already in design centers are designing vehicles for American tastes that are going to then be manufactured at American plants.

50,000 Americans are employed now in Japanese manufacturing plants in the US, producing over two million cars a year. The implication of this is that we really have two auto industries in this country. We have the big three, which is Detroit based and unionized, and then we have the so-called transplants, the foreign plants. And these are not just Japanese. Mercedes, BMW, also are manufacturing cars here. They're not Detroit based. They tend to be primarily in the south. They're all non-union in spite of the efforts of the United Auto Workers to unionize them.

As the big three plants have closed. We have this very unfortunate imbalance of more retirees than active workers. 370,000 retirees and only 290,000 active workers in the big three. The implications of that are severe in terms of legacy costs for health care and pension. It costs the big three about $1,360 a car versus about $80 to $107 a car for Toyota and Honda.

Furthermore, union contracts specify that if workers are laid off, they still get paid. This has led General Motors to continue auto production at a constant level during the recession. The result of that has been an oversupply of cars. The only way to get rid of those cars is to offer these incredible rebates, which you're seeing, $5,000 or more per car, 0% financing. And the other manufacturers have had to respond to compete. This has had a major impact on profitability.

General Motors does not make money selling cars in the United States. General Motors makes money in auto financing and home mortgages, ditech.com, and in selling cars in the Asia-Pacific market. We've seen an unfortunate gradual decline in the market share of the big three. In 1978 they had 84% of the market here in the US. Today, in 2003 61%. The May figures just came out a few days ago and the market share is now below 60%.

In addition to this, there's increased competition from Korea. Hyundai for many years who has had a reputation for very, very poor quality, and now it's rated extremely high from a quality point of view. And Hyundai is the fastest growing car company in the United States.

In addition, US consumers have a much greater choice of the cars to buy. Between 1974 and 2004 there has been a doubling of the different types of models available to US consumers. So we have Volkswagen, the so-called people's car company manufacturing inexpensive cars that now this year for the first time is selling the Phaeton in the US. Prices start at $65,000 and go to $125,000.

If you think that is a lot of money to pay for a car, think again. Daimler, which produces Mercedes, has come out with the Maybach. $350,000 to buy a Maybach. At the same time that Daimler is moving upscale, they are also moving downscale. So next year you will be able to buy the smart vehicle which is currently in Europe. This is a mini car. BMW will be introducing the BMW 1 series.

There are new market segments. Toyota, for example, is offering a hybrid gasoline electric car as is Honda. This is the Prius. It's now second generation. Toyota has found that their market share among youth was dropping, so they've come out with a new product line. Interesting looking vehicle. This is the Skia.

So what's happening is the number of market segments is increasing. Each manufacturer feels the need to cover every market segment. So there's great duplication. They're all chasing market share and there's only so much market share. They're moving upscale. They're moving downscale. You combine this with General Motors overproduction of cars with the new competition from Korea, and there is a huge, a significant worldwide problem of overcapacity. This is a troubled industry and it shows in profitability.

Periodically each one of these companies has been in pretty dire financial straits. You might have seen about two weeks ago that Mitsubishi just got bailed out by one of the Japanese banks. Volkswagen, Fiat, had major problems right now. The only car company, which is in really good financial shape, is Toyota. The market capitalization of Toyota is greater than the combined market capitalization of GM, Ford, and Daimler Chrysler, to give you some idea.

But let's talk about the future and what auto production would look like. And let's start with the car buying process. This is an ad from the Boston Globe that I cut out about a week ago for one of Boston's major car dealers Herb Chambers. And Herb Chambers advertises over 1,000 new vehicles in stock. Think for a moment of the carrying cost of 1,000 vehicles in stock. The land that those 1,000 vehicles sit on. Let's assume that a vehicle goes for $20,000. Not unreasonable assumption. $20 million of inventory is sitting on that dealers lot.

Herb Chambers made estimates, or one might say guesstimates, of what cars people would like, but he doesn't always guess right. So here we are about to see the 2005 cars arrive and we still have brand new 2003 cars left over. So if any of you are interested, Herb is having a blowout sale--

[LAUGHTER]

--to get rid of his 2003s. This excess inventory is symptomatic of major inefficiencies in the car production process. Manufacturing takes only 1.5 days of what is a 40 day process. From the time a dealer places an order with the factory to when that vehicle is delivered takes 40 days. It is essentially an industry push system, where on a quarterly basis each manufacturer determines how many cars of each model is going to be produced. Sometimes they're right. Sometimes they're wrong. Sometimes there are too many. Sometimes there aren't enough.

Instead the industry needs to move to a customer whole operation and a bill to order system. Now, bill to order of course was pioneered by Dell computer and the automobile is much more complicated product than a desktop computer. But nevertheless, Chuck mentioned before the machine that changed the world, which was done by the International Motor Vehicle Program. That program continues here at MIT, and the researchers on that program have estimated and in fact designed a system where it will take five days from the time that a customer orders a car at a dealership to when it's delivered, eliminating the need for not only all this inventory, but numerous inefficiencies in the supply chain.

And it's ironic, but just this week the book Second Century, which is a follow on book to Machine That Changed the World was published by MIT Press. And so some of you may be interested in more details. You should read that book.

Let's talk about automotive markets. The US is basically a replacement market, and Western Europe is very minor growth. The major growth is in the developing countries of the world. The motorization of the third world with a particular focus on China. China has 1.3 billion people, and it's largely an untapped auto market. There are 18.6 vehicles per 1,000 people. You can compare that with 300 to 700 vehicles per 1,000 in the Western world.

There's been explosive auto market growth. 3.25 million vehicles built in 2002. That's a 56% increase. 4.4 million vehicles in 2003. That's a 37% increase. There is improved quality of the cars that are being produced. GM Shanghai reports about 23 bad parts per million delivered compared with 3,000 bad parts per million five years ago. The energy and environmental impacts are severe and several of our speakers today will address those issues.

What is unrecognized, what is really a tragedy, the traffic fatality rate in China-- last year with 24 million cars there were 104 deaths. Now, this largely inexperienced motorists and the large number of bicyclists. And you compare that with Japan, which has three times the number of cars, and the fatalities were 7,000. Even in the US with all the cars that we have, the auto fatality rate is about 45,000, but less than half of what it is in China.

I have grave concern about the auto industry being overly dependent on China. There are of course great economic uncertainties. If China has a hard economic landing, it will have a hugely severe impact on this industry. It's obviously going to affect lots of industries, but this industry in particular.

I was sitting in the front with Norm Augustine and Norm just got back from Saigon and he was saying, boy, you won't believe the traffic there. And you go to Bangkok, as these developing countries are motorizing, there is this huge, huge congestion. And, of course, the congestion is developing as well here in the developed world. It's a very major problem.

What I'd like to do is describe a few very interesting approaches that some cities are using to try and minimize the congestion and how some of those approaches might suggest what the highway system of the future is going to look like. In 1975, Singapore introduced the first congestion pricing system in the world. The idea was if a person wanted to travel into central Singapore certain hours of the day, they had to pay a fee.

Three years ago this system was automated and an electronic road pricing system was put in place. You can see here in the vehicle there's an in vehicle unit where the driver uses a smart car to register to go into the city. There's several checkpoints where the vehicles are checked to make sure that they have registered.

There are two objectives of this system. The first objective is to encourage people who previously drove to switch to mass transit. And it's been very successful. The second objective is to smooth the flow of traffic throughout the day so that the charge varies by each half hour of the day. If Singapore sees that there's a period of the day when very few people use the system, they're going to put a low charge on, and they'll put a high charge on at times when lots of people use it causing people to adjust.

The same sort of principles of yield management in the airline industry and what electric utilities do to smooth the utilization of the resource. Now, this was done in 1975. Since then no other city has implemented congestion pricing because the politicians were afraid. They said, my goodness gracious. What is going to be the backlash from citizens if they're charged to drive into the city?

That is until this past year when Ken Livingstone, the mayor of London, introduced a congestion pricing system in London. It's within this area. Covers about a million people. For those of you not familiar with London, this would correspond roughly to the area in Manhattan from about 59th Street to the tip of Manhattan, the financial district, the east side and the west side. Travelers need to pay five pounds a day-- that's $8. --and they can register by phone or internet or have a ticket bought at a number of shops.

This is not an automated system. There are about 300 surveillance cameras around this area to check licenses and make sure that the drivers are registered. It's been a tremendous success. Everybody was skeptical. It would never work.

Here's some of the data. Traffic delays inside the zone are down 30%. Traffic entering the zone is down 18%; 55% of those people have switched to public transport; 20% have switched to another mode, primarily taxi; and 25% who previously traveled through the zone but had no interest in anything in the zone-- it was simply the shortest route. --now travel around the zone. There's been no adverse traffic impacts outside the zone. There has been a decline in retail sales, but it's unclear whether that's due to congestion charging or whether it's due to the recession, which has affected all retail business in the UK.

Let's now look at another interesting approach with pricing. I think many of you are familiar with the concept of high occupancy vehicle lanes for car pools, van pools, and buses. And this is characteristic of many of those lanes. There's a lot of excess capacity in those lanes. So the idea is to create from those lanes high occupancy toll lanes. So in addition to the multi passenger vehicles traveling free, single occupant drivers will be allowed to travel congestion free in that lane if they pay a fee.

Variable pricing is used to make sure that the lane stays congestion free. So if we start to see it build up with too many cars, the price goes up. The price changes every six minutes. So you know you're going to get a congested free ride.

There are three positive impacts here. First of all, those people willing to pay will get a congestion free ride. And at first they called these Lexus lanes, the idea they're for the very rich. But it's very interesting. They've done surveys and all income groups are using them, and they're not using them on a daily basis. They're using them when they really need to get some place quickly.

So those people travel congested free. You're taking them off the regular highway routes and a lot of money is coming in which can be used to improve the transportation system. There have been successful implementations in California and in Washington DC of these hot lanes. So successful that Maryland and Virginia are now suggesting hot networks be constructed. And what this suggests is that we will be moving to a two tier or dual network of highways offering different levels of service. You will have the existing congested highway system where you travel for free, or if you're willing to pay a fee you can travel on the congested free HOT networks.

Now, both the road pricing system in Singapore and the HOT lanes use electronic toll collection so that a person doesn't have to throw money in every day. They're billed on a monthly basis. This is one application of an Intelligent Transportation System that links together the driver, the vehicle, and the roadway. And this has profound potential impact.

There are other examples. Many of you are familiar with OnStar by GM. We have navigation systems, driver information systems, to tell the driver how to go from point to point and in an emergency to get service to the driver. But the most promising potential is the ability to better manage the entire transportation system, to be able to optimize traffic signals, to control ramp access, to spot accidents when they occur that block lanes and remove those vehicles, to show boards where you can say which parking lot still have space available for drivers, for people waiting at bus stops to say when the next bus is going to arrive. So what we are creating is an information infrastructure to manage and control the physical infrastructure.

The highway system of the 21st century will be built around these concepts of pricing and the use of information and control technology and ITS. We'll have congestion pricing to control congestion primarily in the downtown areas and other congested highways. We will have variable pricing on HOT lanes to offer congestion free travel for a fee. And this will all be under Intelligent Transportation Systems to manage and control the network. So what we will see is a dynamic responsive transportation system that is adaptable to changing traffic conditions and provides choice to the driver as to what they want to do.

Now, there was limited time and I was going to go into another innovation. I'm going to mention it briefly. I'm going to mention it briefly because it was started by MIT graduates. And this is the concept of car sharing. And this is intended for people who live in the metropolitan area to have only occasional use for a car and they don't really want to buy a car because of the difficulty of parking and the cost of maintaining that car.

Zipcar started four years ago. It has 1,600 members and there are 145 cars strategically parked around Boston, Cambridge, and Somerville. In fact, there's a Zipcar in the Kresge parking lot right here. 12% of the members who've joined Zipcar have sold the car upon joining. 25% to 35% have avoided purchasing a car. So 80% of the Zipcar subscribers don't own a car. 95% of them don't use a car to get to work when it's most difficult to use a car.

The estimate is that each Zipcar is replaced seven to 10 cars, and that Zipcar has taken 15 cars off the street. You know how severe the parking situation is in Boston. Taking 1,500 cars off the street is non-trivial. So here's an example of how you're trying to provide choice to urban residents and how you're trying to make the automobile more harmonious with the urban environment.

I've talked about the automotive system as a very complicated system. I'd like to spend just a moment in concluding talking about the implications of this for MIT. It's a large scale system which is characteristic of lots of other large scale systems that we find in the world today, which are increasingly dependent on technology, energy systems, communication systems, the internet. They're large scale networks and they are expanding in size and scope and complexity.

We need engineers who understand technology to design those systems. But those engineers need to know more than just the technology. They have to understand organizational issues. They have to understand issues of the context, the impact, the social, political, institutional factors. So what we need really is a different new type of engineer that works with the traditional engineering scientists, the technical specialists, an engineer who has a broad understanding of how to design large scale complex systems.

And as usual MIT is in the forefront. Five years ago MIT established the Engineering Systems Division in the School of Engineering. But although it's in the School of Engineering, it is a cross cutting unit which has faculty coming not only from the engineering departments, but from the Sloan School and the humanities departments as well.

The faculty and ESD has part of their appointment in ESD and then the other part of their appointment in one of these units. So it's an organization with porous boundaries. It's an interdisciplinary unit focusing on the creation of large scale complex systems. There are two objectives to ESD, broadening engineering education and practice, and developing a new field of study.

In much the way MIT after World War II revolutionized engineering education and practice by introducing the concept of engineering science, so to we believe a similar sort of impact is going to happen with regard to engineering system. Doesn't mean engineering science is less important, but we need both. And MIT is taking the lead role.

So in conclusion, let me just summarize. What I've described is some really fundamental change, some fundamental change with respect to how we might produce cars, fundamental change in where cars are going to be sold, fundamental change in terms of the way our transportation networks are going to operate.

What I haven't talked about is some of the really important harmful side effects of automotive use, in particular the issues of energy utilization and the issues of environmental impact. And I haven't talked about them because of their importance. And because they're so important we're going to have several speakers who focus specifically on those issues to see if we are on the verge of the need for fundamental change. So far the car has been a reasonably adaptable technology that has been able to respond to different regulatory needs in terms of air pollution, global warming, safety, et cetera.

So it's my pleasure at this time to introduce our next speaker, John Heywood who is director of the Center for 21st Century Energy and the director of the Sloan Automotive Laboratory. Let's welcome John Heywood.

[APPLAUSE]

HEYWOOD: Let me follow Dan by saying something about options as we look into the future. I'm going to look at the technology that we've got under development that we could put into our vehicles not in a strict engineering sense, but much more in what other possibilities, a robust view of the possibilities and look particularly at the potential impacts that these might have.

Let me make a few observations to set the context, the basic problem is too many of us use too much stuff too often.

[LAUGHTER]

And it's really important to remember that there are these three components of the consumption problems that go with our transportation system, and the challenge is that they're all growing, all three of them. There's more of us, we drive more, and we use even more stuff. So it's getting worse and worse and worse. And as we look ahead we've got a series of problems that I'm going to say a bit about. Now, at the same time, we have a very effective transportation system. It's a vital part of our economy. It's a vital part of our culture. It's been there. It's been evolving for decades and decades. So of course it's very effective.

And I drove from my house in Luton to here door to door, eight or nine miles, 15 minutes this morning. That's pretty effective. Now you say, well, it's early Saturday morning, yes. And, of course, the problem with our transportation system is really we'd love to have the same system, but only 2/3 of the number of people using it. But that's not possible.

Now, most of the environmental problems that transportation contributes to they're getting worse, while too we're making progress. But I'll talk a bit about the broader scope in just a minute. It's particularly true that they're getting worse fast and in a very problematic way in the developing world where there's neither the sort of organizational infrastructure nor the economic wealth to deal with these problems effectively. And why this is a problem is that our system is so vast that changing its impacts take a very long time.

Now, what are these impacts? Well, here's a list. The ones I'm going to talk about in more detail are the top two, petroleum consumption and the greenhouse gas emissions, the carbon dioxide that results from that. And that's very appropriate. It's perhaps our biggest short, medium term problem, long-term too, but $2 plus a gallon, gasoline is underlining that for us.

But you run your eye down that list, in the developed world we're making progress on air pollution and it's very high on people's priority list. I think the health impact is clear and that's the reason. We're keeping some things at bay. Noise, safety's improving. At least in the developed world it's going definitely in the other direction in our developing countries.

The one that doesn't get much attention is the steady degradation of our ecosystems that comes from both the road structure, the facilities that relate to transportation, and the vehicle impacts themselves, air pollution, water pollution, et cetera. Interestingly, a National Academy panel that I served on a few years ago looked at all of these and picked out the ones they thought were most important in a longer term sense. And that was number two, climate change, and the last one, biodiversity, as a consequence of ecosystem degradation.

So there are lots of problems and they result from this. Too many of us, too much stuff, too often. Now, I'm going to focus on the top two and here at MIT I've been involved with colleagues in a number of studies in our lab for energy and the environment, which is the unit where much of the sustainability issues, energy and environmental issues, are being worked on by multidisciplinary groups. And I put the names up here for a special reason.

We've had some extraordinarily talented graduate students work with us in these problems. Very able and very idealistic young men. If you find Felix AuYeung's name is in the top there, he arrived on my doorstep and I took him on for the project, and he proved an inspiration. He kept us older types young, and we taught him how to do a robust engineering. It was a very good combination. And then he went off to try and save China, which he found discouraging, but he's still just as idealistic as ever.

Now, let me comment a bit on the technologies, enough to give you some sense of what our options are. Mainstream technologies, gasoline engines, transmissions, vehicle components, they get better all the time. Some studies we've done show that typically engines get better about 1% a year. By not much you say, but over 25 years that's a lot of improvement. The only trouble is what we've done the last 20 years is we've traded increased efficiency in all of our vehicles. We've moved up in size and weight and we've demanded higher performance. So we've eaten up all those gains in efficiency.

There are lots of options. Some details. Some bigger changes that are being developed. Some are coming into production. A lot of promise. We can make our technologies better, but there's a cost. It costs a bit more and the demand for fuel economy is low.

Dan mentioned hybrids. I'm sure you've read about hybrids, this combination of an internal combustion engine with a battery and electric motor. Sometimes you use both. Sometimes one. Sometimes the other. They look very promising in terms of a fuel economy's perspective. Toyota's doing very well marketing its Prius. Honda's following.

Interest in hybrids is rapidly rising. They really look like a robust technical solution, but they cost a couple of thousand dollars more and we've still even at $2. Got cheap gasoline. Diesels are going like gangbusters in Europe. In many countries 50% of the new car sales have half diesel. Half of them are diesels. Half of them are gasoline engines. One reason is it's driven by differential taxes. All fuels are expensive, but diesel's less expensive.

The problem in the US is a business case. Can you make a business case when we don't know if these diesels will meet the very strict emission standards that California and the rest of the nation have coming up in a few years? We don't know whether diesels will make it big. We don't know whether hybrids will make it big. But they're good opportunities.

On the more radical end, of course, there's hydrogen and the technology that would use hydrogen fuel cells. Growing research activities here at MIT in these areas. I'm involved and many others are. Potential, yes. Maybe this is what we have to do to get the carbon out of our transportation fuels. But as I list here, it ain't going to be cheap and the technology isn't yet that ready.

And let me illustrate that by looking at, well, how do these changes in technology have impact, and then on what time scale might we expect improvements in what I call mainstream technology and some of these really radical new ideas help us solve these problems? And I'm going to show you a graph which we've produced that looks at-- you put these new technologies, you sell these new vehicles, they go into the fleet, they're used by you and me. What's the cumulative impact over time? Now we understand our fleet.

There's something like 130 million, excuse me, vehicles out there in the United States driving around. We add 15 million or so light trucks and cars each year, the fleet's growing. We drive ever more miles, et cetera, et cetera, et cetera. All that's well understood and we can describe that.

And here's a graph that I'll spend just a minute on. It looks at the United States billions of liters per year. I have no idea what a billion liters per year is. You probably don't either, but it's an unbelievably large amount of gasoline. And then here we go 1970 at this end of the graph out to 2035. I'm willing to go 30 years into the future. I have trouble going further. And obviously this is history. We are right about 2004 now. You can see things have been growing quite fast.

And the top curve, this dark blue one, is no change. We keep doing what we're doing now. Technology gets better, but we demand bigger, heavier vehicles and better performance. So we use it up. This curve is what we think we might achieve if we start to take doing something seriously, and then if we bring hybrids in so that by 2035 half the new vehicle sales are hybrids, we can turn this curve down.

Now, you're probably looking at that and saying, boy, you're not making much difference. That's the nature of the problem. Cars last on average 15 years. It takes a long time to get new technology, better technology, out there to have a major impact. The bottom curve is if we stop fleet growth and stop driving a little more per car each year. And those are only fractional percentage changes, but they compound over 30 years an extraordinarily powerful way. You know that. When you buy a mortgage you know that modest changes in interest rate are going to have a big impact on your monthly payments.

Now, what this says is with this technology turning this curve around and bringing it down is not going to be easy. And remember that the Kyoto Treaty that we have not in this country signed onto said we've got to get down 5% or 6% below this number. We ain't going to make it. This is our opportunity space. If we do everything we can possibly think of, we get to the bottom curve. But that's a bit too optimistic. Human beings don't like spending extra money to constrain what they do so that we improve the total impact situation.

Now, just a couple of minutes on this. Why does it take so long to have impact? And what I've got here is a table where I've laid out the different stages. I've got an exciting new technology. I want to get it into vehicles and have it have impact. First of all, it's got to be competitive. So it's to develop to the point where it's market competitive. And that's the first stage. And what I've shown here is diesels for the United States with good emission controls, we're about a generation, three years or so, away from having competitive technology. We can do it, but it's too expensive.

Here's an advanced gasoline engine highly turbo charged to follow the diesel path and make the engine significantly more efficient. Again, a generation, roughly three or four years, away from being competitive.

The next stage is, okay, Prius' are out in production, these new hybrids, but by the end of this year there'll be a total of 100 odd thousand out there in use worldwide. That's a small number. There's no impact until we're producing them in the many millions. How long does that take to penetrate across new vehicle production? We've got very little case history to use to work that out. But typically at the best it's going to be 10 years. Our estimate for fuel cells is something like 25 years once fuel cells are competitive.

This is stage two. And the third stage is you got to get a big number out there in use if you're going to impact our energy and greenhouse gas situation. So major fleet penetration. Today cars last about 15 years. So 10 years to get a sizable number out there. Probably much longer if we've got to get the hydrogen out there too. So you add all that up and you take a bit off because they overlap. So 20 years for these relatively mainstream technology evolutions. It's 20 to 30 years. I get told off by two groups on this column on the right. Some people say I'm far too optimistic. Some people say I'm much too pessimistic. So obviously I'm about right.

[LAUGHTER]

My final point is if we are going to do something about it, we've got to raise the importance of fuel consumption in the marketplace. And I've had another very talented graduate student working with me an up bend of [INAUDIBLE] under a series of policy measures that reinforce. We've tried to get stricter fuel economy standards through Congress. Doesn't work. We've tried to raise the gasoline tax. Doesn't work. Let's put all these things together so that they add-- there are synergies. --and if we got a diverse set of measures, the political opposition at a minimum gets diffused.

So our strategy is to look at a series of measures. And here's an example. We've completed some studies that look at what the impact of this is likely to be. And as you glance through it, we're going to make the CAFE, Corporate Average Fuel Economy. This is the United States set of regulations that makes the manufacturers meet these average miles per gallon for their vehicles. So we're quite aggressive, but other people in Congress have proposed this in the past few years and the votes are growing.

We can reinforce that with something we call feebates. If you buy a really big heavy vehicle that consumes a lot, you'll pay an extra few thousand dollars as an additional fee, and that will get recycled to those who buy at the other end, a very efficient vehicle. They'll get a rebate. And the way it works out, the rebates are several hundred dollars. The fees could be $1,000 or $2,000.

We bought a system like this already, gas guzzler taxes. They've driven gas guzzlers off the road, and then we couple this with raising gasoline taxes, $0.05 a gallon a year. We don't want to shock the economy too much. We have to do things like this. And then finally we can push the petroleum industry to increasingly bring renewables into the fuel mix that we use.