« Greener Grenades | Main | GMoke Goes (Partially) Off »

Efficiency, Intensity, and Getting from Here to There

efficiencychartxrpt.jpgCarbon dioxide output from the United States will peak and then begin to fall in just a few years, according to the numbers derived by John Whitehead at the Environmental Economics blog. The reason is that carbon intensity -- the amount of carbon produced per dollar of GDP -- is dropping at a rate faster than GDP is growing. At the current pace of intensity reduction, CO2 output in the US will peak in 2008, and begin a gradual decline thereafter. (We previously discussed carbon intensity here.)

This is good news for a number of reasons, not least that it suggests that the current biggest contributor to the greenhouse effect could, with a bit more effort, achieve a far more dramatic reduction in CO2. How to do this is a mainstay of discussion at WorldChanging; here's a look at some of the numbers underlying these options. CO2 intensity is a function of two components: the energy required per dollar equivalent of GDP (or use efficiency of energy); and the CO2 output per MW equivalent of energy (or carbon efficiency of energy). By taking a closer look at the data, we can see which one has mattered more -- and which could stand some improvement.

The material I'm using here comes from the Energy Information Administration at the U.S. Department of Energy, which collects data from around the world for compilation and comparison. The current spreadsheets cover 1980 through 2003; I would expect 2004 data to show up (at the same URLs) by the middle of next year. I am using just a few data sources, but if you have any interest in the subject, the EIA site is a treasure trove. For the record, I recognize that the use of GDP poses problems, since the numbers become dependent upon exchange rates and GDP measurement methods; I would say, however, that this gives us a good estimate, even if the overall accuracy can be questioned.

Carbon intensity is carbon emission per GDP (XLS). The EIA uses slightly different measurements than Dr. Whitehead does on his site, and since I have direct access to the EIA data, I'm going with that. According to the EIA, the United States has seen a gradual but steady reduction in the amount of CO2 produced for every (constant) dollar of GDP. In 1980, the US put out 0.92 metric tons of carbon dioxide for every thousand dollars of GDP; by 1990, that was down to 0.70 tons/$1000; by 2000, it was 0.59 tons/$1000; and the most recent data, from 2003, has it as 0.56 tons/$1000. That's a pretty big drop -- how does it stack up against other countries?

I've picked a few nations that we talk about on a semi-regular basis as comparison: Canada, Brazil, Sweden, the UK, China and Japan. Here's what the numbers look like for those countries:

Carbon Intensity (Tons CO2/$1000 GDP)
Country          1980          1990          2000          2003
US               0.92          0.70          0.59          0.56
CA               1.12          0.89          0.79          0.80
BR               0.50          0.58          0.57          0.55
SW               0.55          0.28          0.24          0.22
UK               0.70          0.53          0.38          0.37
CH               8.46          5.44          2.81          2.58
JP               0.34          0.24          0.25          0.25

Here are two charts of this, with and without China (since it messes up the curve):



A few things stand out: the US and Brazil now have very similar energy intensities; Canada is noticeably less efficient than the US; and China is both incredibly inefficient and getting better at a rapid clip. (India's numbers, if you're curious, are a bit lower than China's but not getting better as quickly.)

We've discussed these numbers before; the usual explanation for the US & Canada is that, as bigger countries, they require more driving, and driving remains heavily carbon-intensive.

Okay, let's look at energy consumption per GDP (XLS), or what I term above "use efficiency of energy." This is in BTUs (a standard energy measurement) per $1000 of GDP.

Use Efficiency of Energy (BTU/$1000 GDP)
Country          1980          1990          2000          2003
US              15,174        11,901        10,081          9521
CA              24,063        20,751        18,131        17,863
BR              10,778        12,481        14,259        13,944
SW              13,175        11,334          9299          8294
UK              10,112          8153          6711          6427
CH             101,936        65,522        35,973        33,175
JP                5508          4450          4703          4605

And, as before, two charts:



These look very similar to the CO2/GDP lines. It's a little surprising just how use-inefficient Canada is; at its current rate, China may well exceed Canada's efficiency by early in the next decade.

It's clear, however, that -- with the exception of Brazil -- there has been an overall improvement in use efficiency. This isn't a surprise, as we've talked about improvements in use efficiency before, but it's good to see real numbers. These improvements come from a combination of factors: appliances and production processes using less power for the same output; changes to practices that reduce the amount of wasted energy; and changes to how we do what we do, whether it's a shift from manufacturing to services or a shift from business travel to videoconferencing. The important point is that improvements remain possible with use efficiency, and most places are getting better at it without putting much effort into the process.

But that's just half the story; improved carbon intensity also comes from reductions in the carbon dioxide we put out when we use energy. The EIA site doesn't have a global spreadsheet of this value, but it does have both overall emissions (XLS) (in megatons of carbon dioxide) and overall energy consumption (XLS) (in quadrillion BTUs, or "quads") -- figuring out CO2/Energy becomes an exercise in fiddling with Excel.

Carbon Efficiency of Energy (MTCO2/Quad)
Country          1980          1990          2000          2003
US               60.7          59.0          58.8          58.7
CA               46.7          43.5          43.5          44.5
BR               46.3          38.5          40.3          39.8
SW               41.9          24.3          25.7          26.8
UK               68.8          63.7          56.9          57.4
CH               83.0          83.1          78.1          77.9
JP               61.6          54.4          52.3          53.8

And the chart, including China since it's not as far off from the rest as in the previous examples:


Unlike the other charts, which generally show a clear improvement since 1980, carbon efficiency appears to have been more-or-less constant over the past 25 years or so. A few of the selected countries do show signs of improvement; Sweden has the most notable change, cutting its CO2 emissions/energy by almost half from 1980 to 1991 -- anyone have a good explanation?

What this chart tells us is not a surprise. By and large, few locations have moved away from CO2 producing energy sources to any significant degree. Clearly, this is the part of the function most in need of improvement; this chart is a way of measuring the impact of the changes we try.

Examples of what could change this chart: adoption of more renewable energy sources; adoption of plug-in hybrids on a large scale; adoption of more nuclear energy; adoption of widespread carbon sequestration projects; peak oil. In general, any new process or practice that reduces the reliance upon high-carbon energy sources in favor of low-carbon sources, or that directly prevents CO2 from getting into the atmosphere, should result in shifts on this chart. If, conversely, we adopt superficially green practices that don't actually make a big difference -- such as some kinds of biofuels -- that will also be visible here.

The last chart to keep in mind tracks use efficiency and carbon efficiency, globally:


This should give us some hope: both drivers of carbon intensity are improving. If we had more time, we would eventually start to see reductions in global CO2 emissions without added intervention -- but time is exactly what we don't have. Both our energy efficiency and our carbon efficiency need to get much better if we're going to make it through this century -- or potentially even the next few decades.

Comments (25)

An interesting article. However, I worry a bit about the statistics. Any indication about the error margins on those figures? I also note that the final graph appears to be leveling off, not improving...

I'm not sure about the margins of error; all of the material is from the DOE website, as linked, so that would be the first place to start looking.

And you're absolutely right -- the most recent few years have seen a global leveling off of efficiency; given that (as shown) some big places are still seeing improvements, it will take some digging through the data to figure out what's causing this flattening.

Mathieu DD:

I would guess that part of Canada's poor performance is due to the colder climate. We would need to compare the efficiency of Scandinavian countries or Russia which have similar weather.

Perhaps a breakdown of the northern states of the US versus the southern states would show some variations there as well.


Given the massive outsourcing of manufacturing to China over the years, we may be understating the efficiency of the U.S. when one considers our true impact based upon what we consume. We may be exporting our carbon intensive industry to China.

So, when cities or even individuals calculate their carbon footprint, consideration needs to be given to consumer goods.

This may be another good reason for a carbon tax so that goods can reflect their carbon costs to society.

Those countries that can produce goods with lower carbon intensity should be rewarded versus those countries who are not reducting carbon intensity.

Btw, Sweden is clearly doing much better than Canada, so there's one comparison showing that cold may not be much of an excuse.


"Sweden has the most notable change, cutting its CO2 emissions/energy by almost half from 1980 to 1991 -- anyone have a good explanation?"


A huge increase in nuclear power, enough to lower fossil fuel use well below 1970 levels while energy demands rose steadily.

Sweden is a very small country and has a lot of hydropower.  Canada might get closer to Swedish figures if it electrified most of its vehicles.


Hydropower is the smallest category of Swedish energy sources. Nuclear and fossil are large, hydro and other renewables are small. Hydro has remained the same since the seventies but nuclear and other renewables have grown. Mostly nuclear.


"Given the massive outsourcing of manufacturing to China over the years, we may be understating the efficiency of the U.S. when one considers our true impact based upon what we consume. We may be exporting our carbon intensive industry to China."

Thanks 't,' this is pretty important... I assume that a big if not disproportionate chunk of China's GDP is from export of manufacturing output, so it would be interesting to divvy up portions of China's carbon emissions by each other nation's contribution to China's GDP through importation of goods made in China.

well, i dont know if you can get away with that as real economics analysis but i think some way to capture the factor of international trade would be pretty interesting. as in, how much are we Americans really "exporting our environmental problems"?

thanks Jamais, for another great post.

The energy per GDP numbers reflect the degree to which an economy depends on energy-intense industry as well as consumer energy use; in particular, countries that are heavy in manufacturing or resource extraction (mining, and particularly coal, oil and gas production) are going to "look bad" from those numbers, but they're really not comparable country-to-country; the worldwide numbers are meaningful though (if the GDP aggregation is valid that is). If you plotted Saudi Arabia, for instance, I'd guess it would look at least as bad as China.

I do wonder about the GDP numbers - Jamais, are those exchange-rate based GDP figures? I don't see GDP on the EIA website, but it's probably buried there somewhere. If so it probably also makes places like China look worse than they should, where the local currency has much greater value on a purchasing-power-parity basis than you would see from exchange rates.

Also are you using the US GDP deflator there? I'm not sure how valid that is worldwide - or even if the way it's currently calculated (since the Clinton administration) is really comparable with previous years; there are some who argue that we have been greatly undercounting inflation (by an extra 2-4% per year) for the last 10 years or so.

The other thing to watch out for in the EIA numbers is the way they count hydropower - not necessarily a big factor overall, but sometimes they give it a factor-of-3 multiplier to be comparable to other primary energy sources, and sometimes they don't. There are probably other quirks of that sort in the numbers.

But nice to see this stuff in graphic form, it's good to put real numbers to arguments, even if those numbers may be a bit shaky...

GDP numbers at EIA

I used the first entry in that list, and yes, they use market exchange rates.

I would love to see how these charts would differ using purchasing-power parity ratings of GDP -- or even a non-GDP metric for development.


If you think that is bas imagine the trainwreck that would have been america joining kyoto.

Not only would all of asia be belching out massively more co2 making all our messy junk BUT we would be bullding nuke plants like crazy already and turning millions of acers of forest land into oil farm croplands for biofuels AND have already deforested all of the rain forests of the world to feed our NEED for so called green fuel.

Oh and likely we would have managed a gloriously worth it ... minor decrease in our co2 levels. While fubaring the entire planet in the process.

The road to hell is paved with good intentions and politicians. The road to heaven is paved with evil intensions and heroes. The road to my home is paved with potholes and roadkill... am I headed to heaven or hell.. I duno but im in for a bumpy ride and its sure to stink.


Good work. Quick point--for clarity's sake, you are showing energy use per dollar of GDP (in year 2000 dollars), not energy use per 2000 dollars of gdp.


Regarding the argument that there is no improvement "because it's been outsourced to China".

Look at the Chinese figures - they are getting better, and fast, not worse.

On the final graph, I think I have an explanation for the slowdown in the mid-90s. A little history; 1985 saw the so-called counter-oil shock, when the Saudis cranked up the taps and the price tanked. If you assume a 10-year lag time for major energy conversion and efficiency projects...that brings you to 1995 or thereabouts, or just the point the curve levels out. I'd predict that the real benefits from 1973 and 1979 oil-shock projects are kicking in on the left-hand side of the graph.

(Thanks for the catch, Drew. Fixed in the text, and I'll fix the graphics when I get a chance.)

Tom L:

This shows the need to reduce more managable GHGs in the short run, and develop technology to manage CO2 in the longer run, as Hansen suggested.


Whether or not China is getting better is beside the point. Outsourcing still results in the U.S. understating its carbon footprint.

Ultimately, carbon emissions are heavily related to the amount we consume and the kinds of things we consume. If there is no impact on the consumer depending on the carbon impacts that he or she causes, changes in behavior will be very difficult to come by.


Well the first thing to remember is everyone lies. So we are much worse off then stated.

Secondly one has to remember that most stats given are wrong even when no one is lieing.

And finaly it must be noted that even when not lieing about the numbers and not being wrong about the stat given much of the time the stat given is not in any way conencted to the real world.

So we have a case where most are poluting more then they say they are and most that do say they are polluting are getting it wrong even when they arnt lieing and finaly we have the simple fact much of whats realy going on in the world isnt even being monitored...

In china HUGE numbers of illegal businesses are operated and no one monitors the emmsions from them and they are very polluting.

In india .. well its india get a bloody clue!

And then we have america.... anyone who thinks that they get the numbers right remember that the same people doing those numbers also think my dad is still alive and that I am 3 different people and a dog and a woman and that im a member of aarp and that im rich.

They also seem to think im ex navy and ex army and that I have a massive trade in adult oriented dolls in asia....


You might want to adjust the GDP figures by purchasing power parity, rather than exchange rates.

Though I must say I don't have a handy time series of PPP over the past 25 years.


could i trouble someone to explain to me and others the goal and implications of "adjusting" the GDP figures by purchasing power parity in this context?
thanks in advance/ sp


Ok in simple terms one measures how much generic moola something is worth the other measures how much generic STUFF its worth.

As the cost of stuff changes around the world and so does the value of moola its good to see each.

Thanks for a great article Jamais. Well done.

Joe Deely:

Great stuff... just imagine where these numbers would be if not for the switch to lower mileage vehicles over the last decade.

A couple of points - you can find some 2004 info at the same web site. For instance - look at http://www.eia.doe.gov/emeu/mer/pdf/pages/sec1_16.pdf for recent data on Energy consumption per dollar of GDP. THe BTUs per $1000 of GDP dropped further to 9320 in 2004. Based on economic growth so far in 2005 and the energy stats through August 2005 - see http://www.eia.doe.gov/emeu/mer/pdf/pages/sec1_7.pdf we should see another significant drop in 2005.

Also, note that the growth rate in world population - another driver of carbon usage - is slowing down. The peak year for population gowth was 1963 at 2.19 %. We are now growing at around 1.15% and this rate is forecast to keep decreasing until it turns negative sometime in the second half of this century.


Mathieu DD wrote "We may be exporting our carbon intensive industry to China."

Also, declining US oil production is being replaced by carbon emitting Alberta oil sands.

tony noerpel:

Just curious if you've thought about the energy costs of necessary stuff like food, transportation, shelter, clothing and say tools. Our GDP today may be inflated by a lot of services and nonessentials like video games and cell phones. The energy required to produce a quantity of food has very likely gone up since 1980. In other words, yeah we can produce more GDP simply by selling our houses back and forth to each other at inflated prices and therefore continue to reduce energy consumption as a function of GDP. But does that mean we are more efficient? If the energy required to produce and transport the things we really need has increased then there is no happy ending in sight.

Allan Ihrer:

Apparently much of Canada's 24% increase in GHG emissions is caused by its developing energy intensive oil sands industry. The input from this industry is supposed to double by 2015.

Link to powerpoint slide with this factoid- http://www.oilsandswatch.org/osf-slide-show/projector.php?slide=11


This page contains a single entry from the blog posted on December 7, 2005 12:36 PM.

The previous post in this blog was Greener Grenades.

The next post in this blog is GMoke Goes (Partially) Off.

Many more can be found on the main index page or by looking through the archives.

Powered by
Movable Type 3.34