Emissions from vehicles have an associated social cost for everyone. We hear a lot about CO2, but CO2 is not the only thing that comes out of a tailpipe. From Wikipedia, emissions components of petrol and diesel engines are:

  • Nitrogen
  • Carbon dioxide
  • Water vapor
  • Oxygen
  • Nitrogen oxides
  • Carbon monoxide
  • Particulate matter
  • Hydrocarbons
  • Sulfur dioxide

My impression is that CO2 gets the most attention because climate change is a hot topic. This doesn't mean that it is the emission component that has the highest cost to society though.

Do any of the components above have higher costs to society than CO2 per same unit of measure (e.g. gram, kilogram, or ton)?

If they have higher costs per unit of measure, do they have a higher cost than CO2 based on their component percentage in typical emissions? The component may be very costly per ton but it makes up a very small amount of typical emissions.

I know that cost is kind of hard to define in these cases. It does seem possible though, because in reports on greenhouse gas a monetary figure is always given.


The US EPA (Environmental Protection Agency) has published information on The Social Cost of Carbon. From the SCC Fact Sheet, the cost social cost of a ton of CO2 in 2015 is between \$12 and \$116 depending on the discount rate used.

The Florida Department of Transportation has also produced computer software that estimates the benefits of transit usage. This is the Trip Reduction Impacts for Mobility Management Strategies (TRIMMS) software. As a part of this, a co-benefit of reduced CO2 emissions is estimated at $50/metric ton. (TRIMMS Supplemental Report page 16)

The costs above are for CO2, but I assume that similar studies have been done for other components of emissions.

  • 2
    $\begingroup$ That depends on many factors, in particular proportions and location. Nitrogen oxide and sufur dioxide cause acid rains that are damaging to monuments of architecture in old cities and poison forests, but don't mean much in desert areas like south-east of the US. Carbon dioxide has a global impact, while carbon oxide is a problem much larger in densely populated areas, while nearly meaningless in rural environments. Comparing the various forms of social impact is a bit of apples-to-oranges. Is it worse if the Colosseum melts, or if a dozen more people suffer stroke in Rome each year? $\endgroup$
    – SF.
    Mar 19, 2015 at 12:01

1 Answer 1


The easy question

Do any of the components above have higher costs to society than CO2 per same unit of measure (e.g. gram, kilogram, or ton)?

Yes, absolutely. This should come as no surprise given the composition of the exhaust—it's mostly inert gases. Water vapor, CO2 and nitrogen gas are all products of the various reactions taking place in the vehicle's engine and exhaust system. An ideal combustion reaction between air and a simple hydrocarbon has water vapor and CO2 as its only products; these are much less reactive than either the hydrocarbon fuel or the oxygen and less reactive usually implies less societal cost.

CO2 gets the attention that it does because of its role in climate change, which is driven by its ubiquity and abundance as a component of anthropogenic emissions. CO2 does not contribute significantly to acid rain; it has no acute* or chronic health impacts; being odorless and colorless, it has no nuisance or aesthetic impact. It's simply a "repeat offender" in terms of society's industrial development because of our reliance on combustion as an energy source.

So, recognizing that the societal costs associated with CO2 emissions are entirely due to its role in climate change, and recognizing that several of the other exhaust constituents contribute to public health, aesthetic and direct economic costs, it is enough to show that any of these has at least as much climate impact as CO2. And that is common knowledge in the field—if you look at this table of direct global warming potentials (GWP) from the IPCC, you'll see that both methane and nitrous oxide (both trace components of vehicle exhaust) have significantly higher GWP** than CO2.

That answers your question but it doesn't tell us anything particularly useful because pollutants in the real world are emitted at different concentrations.

The hard question

If they have higher costs per unit of measure, do they have a higher cost than CO2 based on their component percentage in typical emissions? The component may be very costly per ton but it makes up a very small amount of typical emissions.

Your first question was somewhat trivial to answer; your second is not. There is no typical emissions in a global sense because vehicle engine, fuels and emissions technologies are so varied. There is no typical impact because geography and weather are driving factors. My tentative answer, noting that I do not consider myself an expert in this area, is that we probably don't know. And the reason for this is not that the impacts of these other pollutants are unknown but that the cost of those impacts varies in so many ways, e.g.:

  • The value of lost worker productivity, the cost of health care and access to health care vary hugely throughout the world.
  • Comparing short-term and long-term costs must be done on an equal basis, but what form that equal basis takes is always to some degree arbitrary or subjective.
  • Chronic health impacts are a mess to unravel; there are so many different influences at work over the course of a person's lifetime, and not independently of one another.
  • Aesthetic and nuisance impacts are extremely difficult to quantify; usually reduced to implications for the tourism industry, which ignores the impact to permanent residents.

Translating impact to social cost

Take the example of (ground level) ozone. It's not a component of vehicle exhaust but its precursors are and its concentration is regulated under the National Ambient Air Quality Standards because during the 20th century the U.S. experienced photochemical smog events so severe that residents mistook them for chemical warfare attacks. The main precursors of ground level ozone are NOx and reactive organic gases, also known as volatile organic compounds (VOC) and divided up in various ways under other acronyms that you don't necessarily need to understand, but should be aware of. In any case, these are both components of vehicle exhaust with their own social costs to consider (in the case of NOx, mainly acid rain and acute health impacts; in the case of VOC, mostly chronic health impacts). In combination, they react to form ozone, and this impact is very important—but cannot be attributed to either pollutant alone. Here's a figure showing ozone concentration as a function of NOx and VOC concentrations:


Notice the "VOC Limited" and "NOx Limited" regions of the graph. These indicate conditions where there is already plenty of one precursor, and adding more of that precursor will not appreciably change the resulting ozone concentration.

Now, consider the implications of this for your question. How would you quantify the impact of one ton of NOx? Are you going to assume some ambient VOC concentration? If so, you'll have to work hard to justify it, and it will be highly dependent on location. NOx is mainly anthropogenic*** but VOC comprises many different compounds, some of which are emitted from nice, clean forests. (Thanks a lot, FernGully.)

Is it impossible? No, I could give you a number right now, made of equal parts sunshine dust and horse manure. More importantly, someone—almost certainly not an engineer, note—could spend a whole lot of time and money to do an analysis with some amount of merit and utility that might be compatible enough with the EPA's dollar-value analysis of CO2 emissions to reasonably compare pollutants. To be honest, I don't think there's any need for such an analysis of the other constituents of vehicle exhaust, and if one exists I'm certainly not aware of it.

What's special about CO2?

I know that cost is kind of hard to define in these cases. It does seem possible though, because in reports on greenhouse gas a monetary figure is always given.

Despite the challenges, someone must have at least tried to quantify all these costs—academics need to do research, after all, and where we can identify impacts, we should be able to translate them into costs somehow, right?

In theory, sure. In practice, the scope of the US EPA's economic analysis of the "Social Cost of Carbon" (SCC) is (to my knowledge, again disclaiming the status of "expert") unique. Partly this is because of the huge role of CO2 in both human industry and anthropogenic climate change. Partly it's by necessity, as a result of political, economic and social obstacles—this analysis is a way of justifying legislation, motivating industry self-regulation and influencing consumer choices. Market-based solutions, i.e., putting a dollar cost on everything, are currently in vogue, which drives funding for this type of research.

On the other hand, there are no emission trading programs for uncontroversial air toxics like 1,3-Butadiene, so the sort of research performed on its impacts is not geared toward producing a raw dollar value that could be directly compared to the SCC numbers. Never mind doing the same thing for every VOC species that comes out of your tailpipe.

I'll give you another reason that CO2 is a special case: It's really hard to control. In fact, controlling the other components of emissions usually involves producing more CO2, via more complete combustion in the engine and catalytic conversion in the exhaust system. If we could produce a car that would emit nothing but CO2 and water, that would be a wonderful thing. Meanwhile, not much happens to the CO2 after it's emitted; some of it dissolves into the oceans, but for the most part it persists in the carbon cycle.****

Other air pollutants have been simpler to control and/or less persistent in the atmosphere. Lead has been eliminated from gasoline in most parts of the world (but not all). The social costs therefrom are related to chronic health impacts; these would be calculated (by economists, note—not engineers) by taking into account things like the cost of medical care, what type would be received, the probability of receiving it, the value of lost productivity to the economy—all of this depends on where you are in the world. Similarly, the social costs due to particulate emissions are localized (with one notable exception that I don't believe is factored into the SCC analysis).

Partial answers?

If you limit the scope of your inquiry, there is research out there that you might find helpful. For example, you could take a look at this 2006 analysis of air pollution control regulations in California. It's organized by regulation, and the applicable mobile source regulation is on page 48. The ex ante analysis actually gives some numbers:

The incremental cost of the Rule was estimated at \$65 per vehicle. The cost effectiveness of the proposed regulations was estimated at \$1.20 per pound of NMHC and \$0.10 per pound of CO reduced. ... The cost effectiveness of benzene reduction, attributing the entire cost to the benzene benefit, was from \$4.0 to \$31 million per cancer case reduced.

Keep in mind, this is not the social cost you're seeking, but the predicted net cost of compliance with the particular regulation. And if you keep reading to the ex post facto analysis, you read that the rule "did not lend itself well to ex post cost estimation" because automakers don't want to disclose process and accounting information they consider proprietary and it's impossible to isolate the effects of one regulation in an environment that sees new regulations coming into effect constantly. That's not what makes your question difficult to answer, but I think the challenges are analogous.

And that's after narrowing the scope of the inquiry in several dimensions to a single state, a single regulation, one component of social cost and two categories of pollutant.

An alternative answer

It's important to recognize that passenger vehicle emissions control is one of the major success stories of air quality management (see especially pp. 8-10). As tailpipe emissions get cleaner and cleaner, CO2 would seem to be more and more the pollutant with the highest cost. But it's not as simple as that, because the number of tailpipes is also increasing; note also the predicted changes in fleet composition in Figure 1.

With all this in mind, and thinking to the future, the "easy answer" to your question might be that no sexy dollar-value analysis is required to show that CO2 emissions are the single most important component of motor vehicle emissions by far and will remain so long into the future. They are the hardest for us to deal with, except by radically changing the way we power our vehicles. And that's only a first step; it shifts the carbon emissions to another sector that's just as much in need of transformation, which in turn shifts emissions to manufacturing, etc. The idea is that, rather than simply playing a shell game with the emissions, we realize incremental overall improvements with every move.

And that's the biggest difference between CO2 and these other pollutants. We literally destroy hydrocarbons, NOx and CO in vehicle exhaust systems, which are already very close to 100% efficient, and could always be made more efficient (at greater cost)—but we end up producing CO2 as a result. We can shift away from burning high-sulfur fuels and remove what sulfur we can't avoid—but this process requires energy, and as long as that energy comes from combustion, we're again increasing CO2 emissions.

In a sense, all roads lead to CO2, despite the greater impact of some other exhaust components, pound-for-pound. It's effectively the terminal atmospheric cost of any pollutant we're capable of controlling, and will remain so even as we make improvements in sequestration technology and power generation. That might be ultimately the biggest reason we put the effort into translating its impact into a market cost, when we don't go to the same lengths for other pollutants.

Further Reading

This is already a very long answer, so here's a limited selection of links to resources that I anticipate various types of reader might find useful.

* Within at least an order of magnitude of current atmospheric levels.

** It's worth mentioning here that warming potential is not the only facet of anthropogenic climate change, which also includes, e.g., ocean acidification—mainly driven by CO2. That said, of the climate impacts we are able to predict, warming is both the most well-studied and associated with the greatest societal costs, by far.

*** Non-combustion sources include fertilizer manufacturing and agricultural activity. With some amount of rhetorical contortion, it's possible to consider some of the latter to be non-anthropogenic emissions.

**** The carbon cycle is much more sensitive to anthropogenic emissions than the water cycle; carbon doesn't simply precipitate out as a liquid! A small amount will dissolve into the oceans, but the processes that remove atmospheric carbon are orders of magnitude slower and less effective, as compared to water vapor.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.