CAFE, Part Two

In the same post I discussed yesterday, Keith Hennessey cites the same NHTSA report – the Final Rule governing CAFE standards for model years 2011-15, issued in January 2008 – to make this point: “The proposal will have a trivial effect on global climate change.” (It’s point 5 in his post, and was also picked up by Alex Tabarrok in his endorsement.) Hennessey cites the NHTSA report accurately, but the report itself is misleading.

What does the report say? Look at Table VII-12 on page 624. There are three scenarios that we are concerned with: No Action (which Hennessey calls, and I will call, “Baseline”); Optimized (“Bush Plan”); and Total Costs Equal Total Benefits (“Obama Plan”). If you want to know why Optimized is Bush and TC = TB is Obama, see my previous post. In the year 2100, the projected carbon dioxide (“CO2”) concentration in the atmosphere, in parts per million, is:

  • Baseline: 717.2
  • Bush: 716.2
  • Obama: 715.6

That’s pretty convincing – or is it?

To answer that question, we’ll have to look at the report that those numbers are drawn from: the enormous Final Environmental Impact Statement, whose summary is here. You can see that the two documents tie by looking at Table S-9 on page 14 of the FEIS, which matches the table in the report Hennessey cites.

First, U.S. light vehicles account for a disproportionately large, but still absolutely small share of global greenhouse gas (GHG) emissions. From page 5:

Emissions from the United States accounted for approximately 15 to 20 percent of global GHG emissions in the year 2000. With more than one-quarter of these U.S. emissions due to the combustion of petroleum fuels in the transportation sector, CO2 emissions from the United States transportation sector represent approximately 4 percent of all global GHG emissions. Emissions from passenger cars and light trucks account for about 60 percent of emissions from the U.S. transportation sector.

So the greenhouse gas source targeted by CAFE accounted for 2.4% of global emissions from all sources. If we’re going to look at a 2100 projection, though, we have to take into account the fact that that percentage is certain to go down over the century as countries like China and India become more developed under any scenario. I’m going to guess that by 2100, even in the baseline, U.S. cars and light trucks would fall to 1.6% of global GHG emissions; that means that on average, over the century, they would be producing 2% of global emissions. 

Obviously, if all we do is reduce emissions from one source that accounts for 2% of the total, we’re not going to have a noticeable impact on global warming. If we want to know if we’re doing “enough” in this one little area, the right question to ask is: if we were able to reduce emissions from all sources by the proportion that we are able to reduce U.S. car and light truck emissions, would that have an impact? To answer that question, we have to multiply the impact of the CAFE standards by 50. Then the numbers look like this:

  • Baseline: 717.2
  • Bush (assuming similar reductions of all other sources): 667.2
  • Obama (same assumption): 637.2

That’s starting to look better, but in the baseline, CO2 ppm in 2030 are 455.5, so we’re still really high.

Second, however, the much more misleading part of this report is its assumptions about what happens after 2020.

To begin with, we have to understand what the various scenarios are. It’s not clearly spelled out, but the Summary implies that the baseline scenario is one in which the 2010 CAFE standards are extended to cover the 2011-15 period; see for example footnote 13 on page 13. Most of the other scenarios estimate what would happen with different CAFE standards for model years 2011-15. The important point is that they only measure the impact on model years 2011-15. Relative to the baseline, they don’t even assume that higher standards in 2015 will lead to higher scenarios in later years; they assume that after 2015, CAFE standards revert to the baseline. For example, see footnote 10 on page 8:

NHTSA uses 2060 as the end point for the analysis [of fuel consumption] because it is the time at which 98 percent or more of the operating fleet would be made up of MY 2011-2015 or newer vehicles, thus achieving the maximum fuel savings under this rule.

(Footnote 11 on page 11 says essentially the same thing, in the context of toxic air pollutants.) You see? They are assuming that higher standards for MY 2011-15 will have no effect on standards in later years.

This is absurd. Even if NHTSA never wrote another CAFE standard for any year after 2015, the fact that, under either the Bush or Obama plan, CAFE would be higher in 2015 than in the baseline implies that fuel efficiency will remain higher than in the baseline for some time to come, if not forever. 

For greenhouse gases, however, the analysis is not quite so ridiculous. To see the assumptions, however, we have to leave the summary and go to the Cumulative Impacts section. There, at the bottom of page 53, we see this:


For each alternative, NHTSA assumed that passenger car and light-truck CAFE standards would continue to increase over MY 2016-2020 at their average annual rate of increase over MY 2011-2015. . . . NHTSA assumed further that the fuel economy standards for model year 2020 would remain in effect through the end of the analysis period [2100].

So for greenhouse gases, they are not assuming that fuel efficiency reverts to the 2010 level in 2016. Instead, they assume that fuel efficiency continues to increase in a straight line until 2020 – and then remains flat for the next 80 years. 

This is still absurd, though slightly less so than assuming that fuel efficiency gets worse in 2016.

Graphically, the report is estimating the impact of the alternative scenario (Bush or Obama) versus the baseline like this:


Obviously, it’s true that if you improve fuel efficiency for 10 years, and THEN YOU STOP, you are not going to have a large impact. The point is to continue improving fuel efficiency. And you cannot simply choose your fuel efficiency independently in each year; the achievable CAFE standards for 2021 depend heavily on what the actual standards are for 2020. Put another way, the policy objective is to get on a long-term improvement trajectory that will get us where we need to be, as in the following picture:


The policy whose impact we are trying to estimate is a policy that targets a certain rate of improvement in fuel efficiency, as drawn above – not a policy of improving fuel efficiency for ten years and then stopping, which is what NHTSA estimated. It is true that neither Bush nor Obama proposed setting CAFE standards for the next 91 years, but the rate of improvement over 2011-15 does determine the potential improvement in 2016-20, and the rate of improvement in 2016-20 determines the potential improvement in 2021-25, and so on. To say, as the NHTSA report does, and Hennessey and Tabarrok echo, that increasing CAFE standards for ten model years will not solve global warming is narrowly true but deeply misleading. It’s like telling someone who’s about to go for a short jog, “Don’t bother – once around the block isn’t going to get you ready for a marathon.” Sure, you have to train for months – but you have to start with that first run. And if you don’t go once around the block today, you won’t be able to go twice around the block tomorrow.

This isn’t a Bush vs. Obama thing. Both Bush and Obama said we have to go once around the block today, and twice tomorrow, and so on for a week. (Obama said we should go a bit further each day than Bush, but you’ll note that the difference between Bush and Baseline is bigger than the difference between Obama and Bush.) The issue is that the NHTSA said one week of training wouldn’t be enough to run the marathon. You can’t conclude from that that one week of training is pointless.

The immediate question is, how much bigger is A + B than just A? In the picture above, A + B is 4.8 times as big as A. (Think of A as one slice that is 85 years long on average; B is eight slices, ranging from 75 years long to 5 years long; all the slices are the same width.) This is not quite accurate, first because model years phase into the overall fleet over time, and more importantly because we don’t know how long fuel efficiency increases can be roughly linear. (At some point, if we have practical fuel cell cars, there could be a huge jump upward.) But it will do as a crude ballpark estimate. If we go back and apply a factor of 4.8 to the improvements we saw above (extrapolated over all GHG sources), we get this:


  • Baseline: 717.2
  • Bush: 477.2
  • Obama: 333.2


Now we’re talking.

Of course, these numbers are crude estimates. But this is the order-of-magnitude impact we’re talking about – hundreds of parts per million, not 1.0 or 1.6 parts per million. By extrapolating over all GHG sources, and by assuming continued effort past 2020, we add two orders of magnitude to the NHTSA estimates. These things won’t happen by themselves; we need to do what we can about those other GHG sources, and maintain the political will for more than a few years. But increasing CAFE standards can have the right order-of-magnitude impact; it’s not a waste of effort.

By James Kwak

32 thoughts on “CAFE, Part Two

  1. Take the point where it stands now. CAFE has a negligible effect on GHG. You are tap dancing. Stop hiding behind graphs. In plain words: CAFE pt 2 doesn’t do much.

  2. You know what didn’t actually do anything – going to the moon. Nothing changed here on earth because we brought back a few rocks and put up a plaque except making it possible for college astronomy classes to bounce a lazer off the mirror up there which I must admit is pretty damn cool.

    What did change things was figuring out how to get to the moon. All the engineering, the new materials that had to be developed, the further development of computer techology etc. That changed the world.

    Increaseing CAFE standards to 100 miles a gallon doesn’t do much to offset carbon going into the atmosphere, especiallly with China and India increasing their carbon release far in excess of what even the most ardent daydreaming green could hope for. What it does is require the massive R&D that has gone into increasing horsepower and acceleration over the last 20 years, and force it into fuel efficiency. Maybe nothing will come of it outside of the negligible GHG reduction – but maybe hydrogen, battery, solar technolgies are developed that make a tremendous difference. Combine this with a Cap & Trade policy that creates the same incentives, and stimulus money and the new budget money already allocated to research in this area, and we are setting up a system where we spend hundreds if not thousands of times more money on energy research.

    This is a piece in a puzzle, not everything has to solve every problem alone. What I like about Chu and the Obama approach to this problem is that they understand the technology doesn’t exist to get us off carbon without massive economic consequences. So instead they have faith that we can figure it out, if the proper carrots and sticks are in place.

    It’s really a pattern if you can see it, nudge nudge.

  3. You’ve made a point that needs to be brought up more, namely that simple things like improved CAFE aren’t going to accomplish what we are told needs to be done. That said, it seems that a linear improvement seems as probable as liner non-improvement. Will we (humanity) be up to the changes needed to avert disaster?

  4. “…we need to do what we can about those other GHG sources…”

    Ah, yea.

    Prediction 1: Whatever (minor) reductions in worldwide GHG emissions result from altered U.S. CAFE standards over the next 90 years, it will be SWAMPED by increased emissions (of all types) from other large industrializing nations. And who are we to tell the Chinese to take the more expensive route? Limosine liberalism on a grand scale.

    Prediction 2: It won’t matter, as global warming will remain unproven over the next 90 years, let alone caused by man.

  5. If in 2020 the evidence says we need to keep reducing average mileage and want to bring pressure on the Chinese to also work on reducing mileage it will be a lot easier to get them to participate if we have already been doing something.

    It is part of a path dependency problem. If you start now with minor steps and in 12 years you find that there is not a problem you have not done much damage.

    But in 12 years you find that you need to keep doing something the fact that you have already started make it easier to implement the improvements and get other to go along. Also remember the Chinese and Indian will be a lot richer in 12 years and should be more willing to cooperate because of that increased wealth.

  6. c smith, perhaps you are not aware that China has had stricter fuel economy standards than the US for a while now. Most vehicles sold in the US would be illegal in China.

  7. Yes, I’m aware of that. I’m also aware that China is building coal-fired power plants at a rate which scares Greenpeace members right out of their orange jumpsuits!

  8. I have to agree this seems a little contorted.

    Let’s face it. The new CAFE standards are not a serious policy initiative; they are a photo op. Like most of what Obama has done so far.

    As a conservative (sort of) who voted for the guy and who believes in global warming, I am becoming increasingly disappointed. Higher CAFE standards will actually lower the price of gas and oil. Seems to me reducing carbon emissions ought to entail the opposite.

    We are about 20 years overdue for a carbon tax. (I would accept cap&trade as an alternative. Grudgingly.)

    But this CAFE business is a joke, and not a very good one.

  9. The “trivial effect on global climate change” arguments are going to get a whole lot of play, I fear. You completely lost me very near the beginning, but even my liberal arts brain knows that if your denominator is always global, any single action will be trivial, even really really big ones. Brings up some pretty basic ethical questions, no?

    (By the way, saving 10 million lives impacts only .14% of the world population. Any discouraged cancer researchers out there?)

  10. Yep, way too contorted to get the result you want. You talk about the benefits of concerted actions across all elements of GHG, but not one mention of the costs or route you would take to get there. You can’t have it both ways, either discuss the costs and benefits of just CAFE, or discuss both benefits AND COSTS of CAFE and required miraculous turnarounds in GHG emission.

    Also, if your sole purpose is to argue Obama is awesome (seems to be a recurring theme for you, BTW), why bother assuming via extrapolation that GHG emissions will magically be cut across the board when you can just assume Obama will ascend into the heavens to repair the ozone layer?

  11. If you assume that on going technical progress should be applied to the Bush and Obama projections, some improvement should also be applied to the baseline projection. Its not realistic to assume that it will stay constant. Unless, you believe that at some future date the population will decide that GHG emissions are no longer a cause for concern and lose interest in purchasing efficient cars.

  12. What we really need to work on is controlling water vapor, a far more powerful greenhouse gas, and, therefore, according to the logic of those concerned about anthropogenic global warming, a much more dangerous pollutant. To do this we will need to figure out how to prevent cosmic radiation and solar flaring from effecting the temperature of the oceans, which has a powerful effect on the out gassing of CO2 and water vapor.

    Even better, we need to start now to think about how we are going to deal with the next cycle of global cooling. (In the nineteen seventies, global cooling was the big fear.) We will need to develop plant species that can grow with low temperatures and humidity. We should probably check what the Vikings grew in Greenland during a previous warm period.

    That should provide a whole new cycle of massive government funding for a whole new breed of climate modelers.

  13. I think that as long as it gets the technology moving in the right direction, it’s a good thing. And there are more than just environmental reasons to use less oil. (See Gulf Wars 1 & 2, 9/11, etc). And it’s better to have the government set the standards and let the private sector find the solution than to have the government working on the technology.

    But a fuel tax would work far better. CAFE is too easily manipulated and will take too long to take effect.

  14. not to mention that in a few billion years the sun is going to expand wantonly and engulf the earth. we should nuke it pre-emptively.

  15. “(In the nineteen seventies, global cooling was the big fear.)”

    No, in the 1970’s there were a few nuts on the fringe who were promoting that idiocy. It was NEVER considered to be even a reputable idea, much less a mainstream one. But science deniers seem to love that old chestnut.

  16. The answer to that is not to give up, but rather to do something about our own emissions, and then use import taxes to penalize nations that have not done the same. If China’s exports are at risk, they will stop building coal power plants.

    It would also help for the US and Australia to stop exporting coal; that would drive up the world price a bit, and also affect what China builds.

  17. Wow, so you are going to deny the entire U.S. consuming populace the ability to buy what they want at market prices, as well as denying a huge swath of the existing coal biz its very existence. Not to mention preventing a billion Chinese from becoming wealthier via any route other than the one you deem “clean”! Seems like quite a days work.

  18. Externalities distort market prices and cause markets to do the wrong thing. There is probably no larger externality than failure to price the effects of greenhouse pollution and global warming. Correcting this market failure is important to making economics work for us.

    I doubt the effect of my suggestion would be to deny US consumers anything. Exporters to the US would come on board rather than lose their most lucrative market. That is after all the purpose of the proposal.

  19. Unfortunately, they were the same “nuts on the fringe” that now push the warming fear. Being mainstream doesn’t make it science.

  20. I find it odd that emissions forecasts never consider resource availability. The way I see it, we’re going to burn all of the conventional oil and natural gas — it’s just too good of a fuel source not to. But oil production is peaking and will soon begin to decline (perhaps plateau for a while, if higher prices drive more investment in the extractive industries). There simply isn’t enough conventional oil for the “Business As Usual” growth scenarios of most emissions models.

    The big questions are, what do we do with the unconventionals (tar sands, oil shale) and the coal? If the answer is “burn them”, then we’re going to change the climate rather significantly. A better answer is, “leave them in the ground, and find new ways to generate electricity and power our transportation”.

    Within that framework, I see higher CAFE standards as a good thing. They will drive innovation towards lighter, more efficient vehicles that can run on fewer BTUs, and thus are ready to replace their ICEs with modest-sized electric motors when the time comes. Meanwhile, oil prices will be kept somewhat lower, reducing the incentive to innovate “better” ways to turn unconventionals and coal into liquid transportation fuel.

  21. If CAFE does reduce the price of oil, that will reduce investment and R&D into unconventional sources (tar sands, oil shale). These are not only more expensive to extract, but considerably more carbon-intensive. Instead, innovation will be focused on making lighter, more efficient cars that still meet consumers desires for performance, capacity, comfort, etc. That will make the transition to electric power plants much easier as well.

  22. There are too many unknowns in the climate models to make any sort of prediction on what a given CO2 level will mean for global temperature 100 years down the road. Maybe even unknown unknowns, but the known unknowns are sufficient.

    That is why global temperatures for the past ten years have not followed model predictions. Even over the past 100 years, climate model predictions do not match actual temperatures with a precision approaching the standard statistical significance.

    We are trying to craft policies in a sea of climatological ignorance. The uncertainties outweigh the certainties. I would say the major uncertainty that I have uncovered in my research so far is how cloud cover will respond to rising global temperature. (This assumes global temperature will continue to rise. With the sun’s magnetic field, solar wind, and sunspots declining, the sputtering of our sun may have more impact on global cooling than yet realized.)

    But don’t take my word for it. Check out this analysis on the Nature website:

    And this BBC report offers a good counterpoint — there is no consensus on the IPCC’s level of ignorance:

  23. It cannot be a foregone conclusion that increasing CAFE standards will result in lower consumption of oil. By raising the CAFE standard to 35mpg, you’re telling me that the cost of fuel for me driving will be cut roughly in half for the same number of miles, if my vehicle were currently getting 17 mpg. As people value freedom and mobility, they are just as likely to drive twice as far as before (or nearly so) if the same percentage of their income is now being spent on fuel. A fuel tax would be the answer if you desired to reduce the consumption of oil. Increasing the CAFE standard involves other trade-offs that many may find unpalatable.

  24. I don’t mind carrots too much. I hate sticks. Obama said he would price coal-generated electricity out of the market. For me, that means a future in which I’m huddling around a dim compact fluorescent light bulb freezing my butt off in the Iowa winter, because I can’t afford the power bill.

  25. The “Policy” line on the fuel consumption graph is shown in the vicinity of 100mpg and the text implies that it will continue upward forever. That ignores physics. If cars are hydrocarbon powered, the line must level off. And if cars are powered by something else, then an analysis of the environmental impacts of that something else must be understood before concluding that it is a good thing. The original NTHSA analysis erred on extrapolating to 2100, but not conceptually. The Obama plan imposes high costs and delivers negligible benefit as compared to the Bush plan.

  26. If what you want is R&D, do the R&D, without the enormous nutso side-effects of CAFE. Hennessey and McArdle offer a full catalog of those.

    We need good battery technology – much better than we have today. Batteries are the missing link for the switch to electrics. There are so many other advantages to electrics (no tuneups! no smog checks!) that we would have changed years ago if only the batteries were adequate.

  27. I think you misinterpret this:

    “NHTSA uses 2060 as the end point for the analysis [of fuel consumption] because it is the time at which 98 percent or more of the operating fleet would be made up of MY 2011-2015 or newer vehicles, thus achieving the maximum fuel savings under this rule.”

    It says “MY 2011-2015 OR NEWER” vehicles. Those newer vehicles have to meet the higher standards, too. So there is no “reversion” to the baseline.

    The broader question is what constitutes a fair interpretation of the effects of some policy. “Anything is better than now” is one test, but I think we have to do better. Assuming that your policy gets adopted worldwide is also thinking pretty wishfully. Cost/benefit analysis keeps on rearing its ugly head as the reasonable way to look at stuff like this. So what is the cost of the new CAFE? We don’t know. What is the benefit in economic terms? We don’t know.

    CAFE is the dumbest of the three now under discussion (vs Cap/Trade or carbon tax). It’s a pretty sad commentary that it’s the one (and apparently the only one) that we’re going with.

    One informational question: Can anybody explain the “footprint” bit? That’s the notion that vehicles that cover the same area viewed from directly above (who comes up with this stuff!) will be grouped in some way. Apparently the mfgs don’t get to count “mix” shifts as an improvement. Huh?

Comments are closed.