CAFE Standards are Extremely Inefficient

Or so says Marginal Revolution (via Timmy). I repeat it here to wind up all the people who continually tell me how wonderful fuel efficiency standards are. Prediction: no-one will change their minds.

a gas tax provides immediate, direct incentives for drivers to reduce gasoline use, while the efficiency standards must squeeze the reduction out of new vehicles only. The new standards also encourage more driving, not less.

Context: people keep telling me, whenever the subject comes up "but look how fuel efficiency has increased over the last 2-3 decades, it must be due to fuel efficiency standards" (see comments here for the latest repeats). I invariably reply: "but fuel prices have also increased a lot, how do you disentangle the effects?" And they repeat "Oh, yeah, maybe, but fuel efficiency standards are a really good idea because... obviously" (I may have slightly simplified the discussion, you understand).

This new study also provides something new to the discussion, which I'm sure a moments thought would have provided earlier (so clearly, no-one has been thinking :-): that new efficiency standards only help on new vehicles, which is a small proportion of the fleet, initially; whereas fuel taxes work for everyone.


* Case Overview, CAFE Standards (107th Congress).
* Attack of the Martian Tree Spiders!
* Scientific Meta-Literacy - via P3

More like this

One of the problems with economics as a science is that it is difficult at best, and often impossible, to do a controlled experiment. The economy is also a system with strongly interacting components, so a perturbation analysis of the sort that physicists routinely do frequently gives you wrong results. Couple that with the utter lack of a well-grounded framework, and you see why physics envy can be a major problem in economics.

This example is a perfect illustration. We do know that CAFE standards in the US were gamed: SUVs were officially classified as trucks, even though they are mainly used as cars. That cost of gasoline matters can be seen in why SUVs were adopted: gasoline prices in the US were lower in nominal (let alone inflation-adjusted) terms in the mid to late 1990s than in the early 1980s (the time period when I learned how to drive).

Some of the commenters point out that CAFE standards are at least politically feasible in the US, unlike higher gasoline taxes. There are many reasons for opposing gasoline tax increases in the US, some good (outside of a handful of urban areas, it's between difficult and impossible to be a full participant in the economy without being able to drive, so demand elasticity is smaller than in most comparable countries, and therefore such a tax increase would hammer the rural poor) and some not so good (the Republican Party is opposed to all tax increases, period). Personally, I think the gasoline tax should go up (among other things, to pay for maintaining the transport network we have built), since the tax is per gallon rather than a percentage and therefore hasn't kept up with inflation. But I can understand some of the opposition to such an increase.

By Eric Lund (not verified) on 25 Feb 2013 #permalink

I've wanted gas taxes as the way since 1979 I think. Coulda done it gradually. We'd have a very different landscape today had it been done long ago.
One problem in US: we might have bought more small foreign cars.
I could use spoon feeding on gas taxes in several European countries over time.

"Personally, I think the gasoline tax should go up (among other things, to pay for maintaining the transport network we have built), since the tax is per gallon rather than a percentage and therefore hasn’t kept up with inflation. "

FWIW: The Virginia Legislature has just passed a bill to replace the gas per gallon tax with a 3.5% sales tax on wholesale gasoline sales, with 6% for diesel fuel.

Here's my scientific assessment.

I remember buying gas in 1984 for about $0.90/gallon. "The Inflation Calculator" tells me that is about $2.00 in 2012 money. Yet I filled my tank this morning paying $4.40/gallon. The price of gas in the US has risen a 2x inflation... Does that count as a hefty price signal?

Despite that, SUVs are still popular. Especially among anyone who can make any sort of claim that their SUV is business related. See for example:…

My conclusion: Regulation works! People are still buying gas-hog SUVs because they get to write off the entire purchase amount on their taxes.

Imagine what could happen if the regulations weren't written by the industries they are meant to regulate?

For straight fuel efficiency, you're probably right. For particulates, nitrous oxides, and so forth, emissions standards seem necessary. Particulates and nitrous oxides are the most serious considerations for urban air quality, which is an issue that actually kills people.

[I think I agree on that (I did in the previous thread, too). I can't see any easy way to put price signals on that (there may be an interesting puzzle there for people who *do* want to do everything by price). Regulation seems to be the best way. Fortunately, that area doesn't seem to have strong manufacturer resistance to regulation -W]

There's also the interesting point that higher taxes mask price signals in the underlying commodity, but I'm not sure how important that is...

[See recent UK calls for lower fuel duty due to oil price rises -W]

I'm intrigued to know where these costs of achieving CAFE regulations come from. I don't have access to the paper and the summary gives no detail. I don't know about the US but in Europe new cars have been steadily getting cheaper in real terms in spite of legislators steadily ramping up safety, environmental and noise regulations in the past couple of decades.

[Modelling. I don't think there can be any other way to predict the future. I haven't read the paper itself, either -W]

A 20% reduction in fuel consumption by 2050 in the US is really rather modest ambition and can be readily achieved with off the shelf technology and without any need for alternative drivetrains or fuels - just good engine and vehicle design.

But as I said in the other post have CAFE and a gas tax. Keep pushing the manufacturers to release efficient new vehicles and raise fuel duty to continue the reduction in demand for travel in the US that started in 2008.

[have CAFE and a gas tax - ah, then you haven't read the summary either. That points out that doing both is more expensive -W]

Summary says "Karplus also showed that combining the FES and RFS regulations led to roughly additive costs but less than additive benefits. " but I'm talking about a combination of an FES and gas tax.

Also don't cost-benefit analyses treat taxes as a transfer with no cost to society? If so what is the scale of the costs that we're talking about? Anyone got access and read the paper know?

@Rick FWIW: The Virginia Legislature has just passed a bill to replace the gas per gallon tax with a 3.5% sales tax on wholesale gasoline sales, with 6% for diesel fuel.

VA has a couple of heavily traveled interstate highways which move truck traffic from east to west and north to south. No surprise there.

By Eli Rabett (not verified) on 26 Feb 2013 #permalink

The answer to Tim Worstall's question about why the US has CAFE standards and not an increase in gas taxes is Tim Worstall. Timmy, of course is one of the libertarian crowd who thinks taxes are evil.

[You made that up. It isn't true. Timmy is, for example, in favour of a carbon tax -W]

This is much like Lizzy Bordan killing her parents and pleading for mercy as an orphan.

By Eli Rabett (not verified) on 26 Feb 2013 #permalink

Gator @5: It's even more extreme than that. I was buying gas at nominal prices of under $1 per US gallon in the 1998-1999 time frame (a lower nominal price than I remember from 1984). Nominal prices first crossed $2/gallon in 2004 and $4/gallon in 2008. IIRC the tax writeoff on SUVs for business use was enacted, or at least substantially increased, during that period. Not coincidentally, the then president and vice president had backgrounds in the oil industry.

However, some price signals do work. When I do a road trip to Montreal (about a five hour drive from where I live), I fill my tank shortly before crossing the border. IME gas costs about 30% more in Quebec's Eastern Townships than in Vermont's Northeast Kingdom. I presume most of that difference is due to higher taxes in Canada.

By Eric Lund (not verified) on 26 Feb 2013 #permalink

Wait, a more appropriate response:…
Pull the other one, it's got bells on.

[Read it more carefully. I haven't checked the numbers, but in the UK, he is likely correct. Remember that UK fuel taxes are not carbon taxes, they are just for revenue raising. You should not assume that they are *lower* than a pure carbon tax would be -W]

By Hank Roberts (not verified) on 26 Feb 2013 #permalink

> don't know what that means

That's why I use the apostrophe. I've seen capital L, lower case l, "Royalist" and "Liberal" and "Conservative" varieties -- and then cross state lines and the subtle variation in meanings change again.

Seems they all say their program will be so clearly right that people will rush to the marketplace to buy it, but they haven't got the details worked out quite yet and nobody's buying.

I guess I'm a Leopoldian

By Hank Roberts (not verified) on 26 Feb 2013 #permalink

California introduced car standards that were much tougher than US required. Result is that car efficiency in US improved to meet the California standard.

Should this be portrayed as a victory for standards?

As a way of doing things, this can be argued as inefficient because you get multiple bites at the cherry. So it is inefficient but is that inefficiency actually a sensible approach where fossil fuel money might corrupt the political process.

So is there an argument for saying this approach is inefficient but that inefficiency is necessary and appropriate to counteract money corrupting the political process?

(Doubt I will make any republican friends by posting this.)

I've worked with the model used in this paper, and while it is state-of-the-art, I think it still suffers from the standard-economist-blindness of not taking into account market inefficiencies other than pollution externalities: in this case, I think 1) there is evidence that people apply discount rates to fuel efficiencies in their purchases of new vehicles that are much higher than the "socially optimal" discount rate, and 2) standards can be technology driving. I would agree that given only one of tax or CAFE, I'd choose tax, but given the option to do both... I might do so. (I might also consider a "fuel efficiency" tax: eg, tax fuel, but also tax the sale of new vehicles based on their efficiency).

I've always seen UK fuel duty as something of an argument against carbon taxes, or at least carbon taxes as a way of achieving large CO2 reductions.

Why? Well, even though the UK has effectively a 'carbon tax' on fuel of at least 200%, per capita, milage adjusted usage is still about half that of the US. The implication is that carbon taxes would need to reach extreme levels on other fuels (cf natural gas and FF-electricity) before usage came down much.

This is not entirely unexpected, because of the big upfront investments involved in burning fossil fuels in the first place. CO2 taxes would have to reach very high levels indeed to force a 5-year-old coal power station to switch off and the owners build something else.

Hence a CO2-tax ends up becoming a super-VAT, which ends up being a regressive tax. No wonder timmy likes it..

[I think you too have missed the point - or at least, you haven't addressed it. If you regard GW as an economic problem, in which the costs (SLR, dieback) must be borne by those receiving the benefits (fuel use) then the "problem" you're describing isn't a problem.

If, however, you have moral objections to GW for some reason, then the economic argument doesn't apply -W]

By Andrew Dodds (not verified) on 27 Feb 2013 #permalink

If you regard GW as an economic problem, in which the costs (SLR, dieback) must be borne by those receiving the benefits (fuel use) then the "problem" you're describing isn't a problem.

Ah, well, I think I see the problem then... GW is a problem where the costs are not borne in anything like proportion to the benefits received. Poor Bangladeshis will suffer far more from SLR than SUV owners in the Home Counties. It is a problem where those who cause the greatest proportion are simultaneously the most able to insulate themselves from the costs, simply by virtue of wealth. It's an "economic problem" all right, but not in the way you mean...

[That's a fair comment. What I was trying to do was to clarify the fundamental issues.

So: at the moment, of course, the externalities are only intermittently internalised (e.g. fuel tax in the UK, not in the US, and the madness of fuel subsidies elsewhere). Also, of course, fuel tax receipts flow to the UK treasury, not the non-existent world govt. So while the "polluter pays" in this case, the cost-bearer doesn't get the benefit (and since the polluter gets their share of UK tax receipts, arguably they aren't really paying. That's another matter).

I've touched on this before ( and said "To which I’d reply: well, tough, you can’t solve every problem." I'd still go for that as an answer, now. Because, now, the costs of GW are small -W]

In fact, we don't even need to invoke GW to realise the problems with that line of reasoning, we only need to observe that those who suffer the most from urban air pollution tend to be poor (and the poor are less likely to own cars), whereas the people who drive the most polluting vehicles tend to be rich, and tend not to live in the cities they pollute (because they can afford both to move away and the costs of daily commuting). I don't own a car (and never have), yet I live in one of the busiest and most polluted areas of my city. I bear the costs, others enjoy the benefits.

Because, now, the costs of GW are small -W

Perhaps... But as I observed earlier, there's more to vehicle emissions (and other combustion emissions for that matter) that GW - there's also the small matter of an estimated 13,000 premature deaths a year from UK combustion sources:…

[Well, we can switch topics if you like, though it may be confusing. As I said somewhat earlier, it seems likely that regulation is a better solution to particulate emissions, etc., than fuel taxes -W]

Ok, I was one of those in favour of fuel regs so I'll bite.

I've read the article. The modelling and assumptions behind it are way beyond my comprehension. I am aware that this is what climate septics routinely say before dissing climate models, but (you knew that was coming right?) ...

- A fuel tax is regressive. Low income households spend more of their income on fuel

[But the really poor don't own a car at all. However I'm not prepared to accept the regressive stuff as a killer argument, its just one amongst many, and not very high up. If its the only one you're paying attention to, then yes you'll be opposed to fuel taxes. Perhaps you'd like fuel subsidies for the poor? -W]

The lead author covers this in her thesis (on which the paper is based), and I cherry picked this

"The lowest income category (< $25,000 per year) shows the most inelastic behavior, with the elasticity of demand for VMT not significantly different from zero. Since these households are likely the most cash-constrained, reducing fuel consumption without reducing VMT may be a particularly attractive option, especially if the proportion of vehicle-miles devoted to more essential (non-discretionary) household activities (such as grocery shopping or commuting) is greater than for higher income households. As income increases, household responses generally become more elastic, consistent with the notion that reducing the number and length of vehicle trips may be easier if those marginal miles are devoted to vacation or other discretionary trips, which can be reduced without affecting income, or for which less expensive substitute modes of transport may be available."

Lower income households have virtually zero change in their miles driven when price is increased. They rely on learning more efficient driving techniques or purchasing a more fuel efficient car. Maybe I missed it, but I didn't see in that study where they modelled the efficiencies of the vehicle fleet with and without a FES.

- Their analysis is sensitive to increasing availability of low(no) carbon vehicle fuels (biofuels) as emissions standards do not provide as great an incentive as a fuel tax to replace fossil fuels with biofuels. And we all know what sensible think of biofuels…

You were right, I haven't changed my mind. My guess is neither have you.


If they question is 'who benefits' then it can be even more complex.

If I drive to work, (given that in North Somerset it's the only realistic option), am I benefitting or is my employer? They would have to pay much, much higher wages to eliminate driving to work, so they benefit from my driving as well. And driving is the simplest case.

[No, I don't think this makes sense. You decide where to live, and who to work for, and at what cost, based on various factors. Your employer decides where to base themselves, and what wages to pay, ditto. You're responsible for getting to work; you're responsible for whatever emissions you emit along the way -W]

Electricity? Am I the polluter for using it, or the electric company for generating it? In theory I have the option of avoiding any carbon tax on electricity by adding battery backup to my solar panels.. but the cost would be extreme.

[Again, you -W]

Home heating? Well, I suppose that's more clear cut. Only problem is that I have even fewer options than with electricity.

So, a carbon tax is not as straightforward as the simple minded economists of the world would have us believe. It would certainly be highly regressive - not only to the poor already spend proportionately more on energy to start with, but they by definition cannot afford the avoidance measures (Solar panels, hyperinsulation, EVs, ground source heating, whatever) that is ultimately what we want to see from a carbon tax.

In short: 'Polluter pays' is fine assuming you know who the polluter ultimately IS, and that the polluter has options. Otherwise you are just shifting taxation from rich to poor.

Ultimately, in the UK, carbon emissions effectively come from a small number of sources - Power stations, oil refineries and the natural gas network. If these can be replaced with zero-carbon alternatives - and I see no physical or technical reason why not - then the problem goes away.

By Andrew Dodds (not verified) on 28 Feb 2013 #permalink

Sorry, but to say that I am entirely responsible for my emissions *when I have no realistic alternative* (not work? raise a family in a 1 bed flat?) is very disingenuous.

For electricity it's even worse. I pay the tax but get absolutely no say in future generation capacity decisions.

The logic is very neoliberal in nature.. no doubt to be followed by the classic economists' argument: 'People are choosing to pay carbon taxes instead of choosing other options' (when those other options don't actually exist). A really low blow would be, of course, to point out that at least a fraction of the UK's 24,000 excess winter deaths are due to people making the rational economic choice not to turn the heating up.

So.. so summarize..

- Empirical evidence* suggests that Carbon taxes would have to reach high levels (>50% of current retail prices at a minimum) to actually change consumption.
- Since the people making investment decisions that actually affect FF usage are not the ones paying the tax, market signals will in any case be ineffective.
- A wealthy minority may be able to avoid the taxes by changing where they get their energy from. The bulk of people won't. This makes it an extremely regressive tax. Of course, you could try and make it revenue-neutral.. but that would involve some serious progressive taxation in response. Which would offset people even noticing the carbon tax..
- Allowing for this, I'm sure you'll have raised enough cash to pay for relocating London et al when the WAIS unexpectedly turns into an iceberg circa 2060. Unfortunately you'll have already spent it.

Still, I'm sure it works in an Economics textbook. And we all know how reliable those books are.

* Best I can find from Teh Googles in 3 minutes.…

By Andrew Dodds (not verified) on 28 Feb 2013 #permalink

Most of Timmy's nonsense depends on the multiplier he uses for Jevon's paradox. For the CAFE standards it looks like there is a gross gain of 7% in US energy usage and a 2% bounce back for a net of 5%

By Eli Rabett (not verified) on 28 Feb 2013 #permalink

Perhaps you'd like fuel subsidies for the poor? -W

An interesting question... At the risk of going even further off-topic, what's your position on the Winter Fuel Payment?

Eli's mention of Jevon's Paradox throws up another interesting question... Perhaps we should lower the utility of fuel use by deliberately reducing efficiency? In economics, anything is possible... ;)

>> Electricity? Am I the polluter for using it, or
>> the electric company for generating it?
> [Again, you -W]

Also true for sulfate/acid rain/mercury etc.
Point sources provide economic leverage
that's lost if you wait til the problem's widely diluted.

By Hank Roberts (not verified) on 28 Feb 2013 #permalink

"But the really poor don't own a car at all. However I'm not prepared to accept the regressive stuff as a killer argument, its just one amongst many, and not very high up. If its the only one you're paying attention to, then yes you'll be opposed to fuel taxes. Perhaps you'd like fuel subsidies for the poor? -W"

I think we need a carbon tax (we currently have one here in Australia on the x largest emitters (where 10 < x < 30 -- I can't recall exactly) and it seems to be doing the trick. Importantly ALL the revenue from the carbon tax was used to pay compensation to lower income groups and "export sensitive industries".

Where we differed was I thought FES was a "no regret" option. This modelling certainly seems to suggest otherwise. All I can say is I have no expertise to judge the validity of the result that it is six times more expensive to have a carbon price and FES than a carbon price alone. They do say alot of this expense is due to FES reducing uptake of biofuels, which are contentious in any case.

---Perhaps you’d like fuel subsidies for the poor? -W

Somewhat better than watching them freeze, but you libertarians think of that as a benefit, not a cost.

By Eli Rabett (not verified) on 28 Feb 2013 #permalink

> #32
lost response? Interesting to 'oogle' the issue of subsidizing heating costs, which apparently is much discussed in Britan right now.

[Did it need a response? There's a bit of fuss about winter fuel subsidies being given to pensioners in Spain. That's obviously stupid; the entire fuel subsidies thing is stupid, less obviously so (its not really a fuel subsidy, its just money) -W]

By Hank Roberts (not verified) on 03 Mar 2013 #permalink


I managed to track down a copy of the paper online and see that the total costs for the gasoline tax are given as $0.7 billion pa, FES-sharp is given as $10bn pa and FES-gradual is given as $63bn pa.

Given that the US currently consumes 134 billion gallons of gasoline each year, a $1 a gallon tax would levy $134 billion each year. But as this is a tax it's counted in economics assessments as a transfer so doesn't appear in the cost-effectiveness calculations. As far as the US people are going to be concerned, $134 billion per year is going to be seen as quite a hefty cost and $10 billion per year will seem really rather cheap in comparison.

[Hmm, but I hope you agree that the tax isn't really a cost, it really is an xfer -W]

> [Hmm, but I hope you agree that the tax isn't really a cost, it really is an xfer -W]

Yes I agree it's a transfer. But in the real world of selling a policy to a population, good luck in persuading them that it's just a transfer and persuading them that the government extracting $134 billion from everyone's pockets to elicit the same effect as vehicle manufacturers extracting $10 billion from the 15 or so million who purchase new cars in a given year is more efficient!

[How can you hope for efficiency when the entire population is as incompetent as you describe? -W]

James Buchanan's words seem particularly relevant to this discussion...

"Economists should cease proffering policy advice as if they were employed by a benevolent despot, and they should look to the structure within which political decisions are made. ... I called upon my fellow economists to postulate some model of the state, of politics, before proceeding to analyse the effects of alternative policy measures. I urged economists to look at the "constitution of economic polity," to examine the rules, the constraints within which political agents act.

By Marlowe Johnson (not verified) on 05 Mar 2013 #permalink

> [How can you hope for efficiency when the entire population is as incompetent as you describe? -W]

??? Err what?

[Just up above you've agree that the tax is a transfer not a cost. Then you've gone back to asserting that Joe Public won't see it that way. Ie, you're calling them incompetent -W]