« More About Population | Main | The Seas of Elysium »


earthmeasure.jpgHow should one measure a nation's energy footprint across time? After all, if we're trying to increase efficiency of energy use, we need a baseline to tell how well we're doing. The question takes on a new twist when one wishes to compare two or more places. Simply counting BTUs or CO2 output won't do it; a large country, no matter how efficient, will inevitably consume more power and produce more greenhouse gases than a small one. As it turns out, there are two broadly accepted approaches to solving this dilemma, measuring per capita and measuring per productivity -- the latter generally called "intensity." They each have their advocates: per capita shows consumption per person, a crude version of the footprint; per GDP shows consumption for economic activity, a rough measure of efficiency.

The US Department of Energy Energy Information Administration makes available a huge variety of global energy and carbon measurements. The data are generally available as Excel spreadsheets, and typically cover 1980 to 2002, with entries for most countries. In the extended entry, you'll find links to some key datasets on energy consumption and carbon emissions, as well as an exploration of what these numbers might tell us.

(Unfortunately, only US data is made available in HTML or text format. OpenOffice, available for most platforms (I suggest the NeoOffice version for Macs), should open these Excel files just fine if you don't have MS Office available.)

Looking at the energy and carbon data using the different metrics of per capita and per GDP gives us some interesting contrasts. Some of the results may not be what you might expect. Take a comparison between the US and Canada, for example, over the final five years of the data:

	US million BTUs per capita	Canada million BTUs per capita
1998		352.1				406.5
1999		354.9				416.5
2000		350.6				419.9
2001		337.9				411.7
2002		339.1				417.8

US BTUs per 1995 dollar GDP Canada BTUs per 1995 dollar GDP 1998 11,403 18,959 1999 11,140 18,755 2000 10,977 18,204 2001 10,660 17,508 2002 10,575 17,341

(BTU=British Thermal Unit, the standard non-metric measure of energy.)

Energy use per person and per dollar of GDP is actually higher in Canada than in the US. That is to say, Canada appears to be overall less efficient in its consumption of energy than is the US, or at least was through 2002. That surprised me. Let's look at a few other comparisons.

Million BTUs per capita
	UK	China	India	Japan
1998	167.2	29.6	12.4	168.7
1999	166.5	29.2	12.7	170.8
2000	164.5	30.5	13.3	172.6
2001	166.6	31.8	13.4	172.4
2002	162.2	33.3	13.3	172.3

BTUs per 1995 dollar GDP UK China India Japan 1998 7,860 41,171 27,930 3,880 1999 7,669 38,352 27,306 3,931 2000 7,375 37,293 27,776 3,871 2001 7,336 36,578 27,053 3,858 2002 7,039 35,764 26,198 3,876

From an energy user per-person standpoint, the UK and Japan are close, and both significantly more efficient per-person than the US or Canada; China and India, while even lower per person, are both still developing -- the per capita numbers strongly trend up as more of their populace get cars, connected to the grid, etc. But it's the intensity measure, BTUs per dollar of GDP, that really stand out. China and India burn a lot of energy in their economies, whereas the UK and Japan use significantly less energy per dollar produced than either the US or Canada.

It's hard to draw particular conclusions from this. There may be a geographical factor at work, with physically larger countries requiring more power to move goods around.

But from a climate disruption perspective, energy use is a surrogate for the real problem, greenhouse gas emissions. Let's see how our six nations fare per capita and per GDP.

Metric tons of CO2 per capita
	US	Canada	UK	China	India	Japan
1998	20.63	9.15	9.39	2.35	0.91	8.66
1999	20.73	9.69	9.14	2.30	0.93	8.96
2000	20.62	9.92	9.46	2.37	0.98	9.31
2001	20.04	10.16	9.68	2.47	0.98	9.28
2002	19.97	10.24	9.36	2.57	0.98	9.25

Metric tons of CO2 per thousand 1995 dollar GDP US Canada UK China India Japan 1998 0.67 0.83 0.44 3.28 2.06 0.20 1999 0.65 0.81 0.42 3.02 2.00 0.21 2000 0.65 0.81 0.42 2.90 2.05 0.21 2001 0.63 0.79 0.43 2.84 1.97 0.21 2002 0.62 0.79 0.41 2.75 1.92 0.21

Okay, what is this showing us?

Clearly energy use in the developing world for economic production is much more carbon-intensive than in the developed world. That doesn't tell us anything new, but it does give us some numbers to underscore observations. Per-person, about 10 metric tons of CO2 seems typical for Western-style development. Carbon dioxide output seems to track relatively well with energy use (of course, digging through the spreadsheets, you'll find exceptions -- France, for example, has low carbon output relative to its energy use, arising from its dependence on nuclear power). We find one notable exception to this pattern, however: US CO2 per capita. Whereas American per-person energy use trails Canada somewhat, its per-person carbon load is more than twice that of Canada. That is to say, its carbon output per energy consumption is unusually high for a developed country. What could be going on here?

My first guess is that this is a reflection of American suburban sprawl coupled with relatively low availability and use of mass transit. Low density development requires more driving for the same level of economic output. This seems to be a measure of the inefficiency of sprawl from a carbon output perspective. I'm open to other suggestions, though. A look at the DOE spreadsheet breaking down energy consumption by fuel type might prove useful. Why else might US carbon/person be so out of line with the rest of the West?

The larger question here is what are the metrics we're missing? GDP is a poor measure of actual economic activity, especially when compared using exchange rates instead of the more accurate but harder to pin down purchasing price parity. What would a spreadsheet of energy use per Gross Domestic Happiness show us? Can the metrics linked to in Indicators Are Reality give us a new perspective?

We have abundant data. What we need now is to figure out what the right questions are to ask.

Comments (13)

Canada is very, very, very cold, and a lot of the country has very dispersed populations with long drives - or plane trips - between locales.

I suspect that explains the higher numbers.

Great subject to cover, Jamais.

Energy use per person and per dollar of GDP is actually higher in Canada than in the US. That is to say, Canada appears to be overall less efficient in its consumption of energy than is the US, or at least was through 2002. That surprised me.

That's almost certainly because Canada's colder. Hawaii is probably the "most efficient" state in the US in terms of BTUs/person, but it's not because they do anything magical with technology or design their cities better etc. My state of Minnesota, for exampe, which has huge temperature variations, consumes 52% more energy per capita than Hawaii, which has a very narrow and temperate annual temperature band.

As for the higher carbon output of the US, I would guess that's because coal is the fuel for about half of our electricity, whereas Canada probably uses much less because of better relative hydro resources. Probably the same with Europe because of much heavier use of nukes.

And since I wanted to be sure, I put together a summary spreadsheet of primary energy consumption for the year 2002 for different fuel types for the US, Canada, Western Europe, and Japan, all expressed in BTUs instead of their normal units (eg, millions of barrels per day for petroleum).

The spreadsheet is here.

The source data is from the EIA - here.

You can see that net hydro consumption accounts for only 3% of primary energy consumption in the United States, whereas it's a whopping 24% of Canada's - an eight-fold difference in relative consumption. Conversely, coal consumption accounts for 13% of primary energy consumption in Canada and 23% in the United States - a nearly two-fold difference. So that's probably why you see such a difference in CO2 ouput per capita, not because of spawl/lack of density. Outside of the main urban cores in Canada, you see pretty much the same development patterns as you see here. Get 10 kilometers north of the lake in Toronto, for example, and it's not too different from an American suburban landscape - big roads, strip malls, and the like.

You can see this also in petroleum consumption figures. Counterintuitively, Japan has the highest relative share of primary energy consumption coming from petroleum, even though it has the best public transit system of the three countries and Western Europe. This is most likely due to Japan being so efficient with other energy sources (eg, end use electricity consumption efficiency), such that petroleum becomes a big part of the picture. It certainly has nothing to do with their auto dependence (lack of it), density (lots of it) and sprawl (little of it).

It takes a while to get a sense of what all the data indicate, though I think the BTU/$ of GDP measurement is probably the most useful for rating a country's relative efficiency, as long as you make some adjustments for climatic differences.

Garry Peterson:

Canada is cold and big, but so are Sweden and Norway. And they have a CO2/capita is almost 1/2 that of Canada and the USA.

Canada is cold and big, but so are Sweden and Norway. And they have a CO2/capita is almost 1/2 that of Canada and the USA.

Perhaps I was unclear. Canada uses more energy per capita than the US probably because it's colder. The US produces more CO2 per capita than Canada because it consumes far more fossil fuels as a percentage of total energy consumption.

Sweden and Norway may be similar in climate to Canada, but they consume FAR less fossil fuels as a percentage of their total energy consumption than Canada or the United States -- Norway because of hydro and Sweden because of hydro and nuke.

I've updated my spreadsheet (here) to include those two countries as well.

The point I'm trying to make is that these are mutifactoral questions when looking at national energy consumption figures. You need to consider:

* end-use efficiency
* climate, including degree of variation relative to a temperate level
* distribution of population relative to climate
* distribution of population relative to one another (density)
* degree of industrial activities and economic development
* type of industrial activities
* the transportation system

As for CO2 output, you not only need to look at the above factors, but also the percentage of fossil fuels relative to all primary fuels in overall consumption.

Coal use in particular is dependent on relatively proximate availability of it, as it is heavy and relatively low value - meaning, it loses a great deal of its cost-effectiveness the further you get away from where it is mined. Nukes, for example, don't have this problem.

I too agree that this is a great subject to cover... in fact it may become THE subject to cover in future years if global warming progresses.

If you want to add 2003 and almost 2004 United States numbers to your spreadsheet you can get latest data from http://www.eia.doe.gov/emeu/mer/pdf/pages/sec1_3.pdf

Heres what I see is interesting in the US numbers.
In 2000 we were at 98.905 quadrillion BTU and we dropped over the next couple of years before rising back to 98.733 in 2003. If I exrapolate for the month of December it looks like we will just barely go over 100 quadrillion for 2004. What's interesting about all these numbers??
During the last four years we have added 13 million people, our economy has grown by over 10%, millions of SUVs have been added to our roads but our energy usage only went about a little over 1%!
This is incredible!

Over time the energy intensity for the world has increased by about 1.1% per year. If we could increase this rate, and combine this with a decreasing population growth rate see
then we could bring the living standards of the world up to the current US rate AND keep energy usage stable or even declining. Obviously easier said than done.. but when I look around me and see the inefficiencies I think this is possible.

If we can keep energy usage stable then it becomes
way easier to lower CO2 emissions.

Great topic, and great comments.
I would add that one reason why the US might look like its getting more energy efficient is that all the heavy industry and much of the light production has been outsourced overseas. This has also happened in Europe. This might partly explain why Japan has such high oil use by comparison, because they still actually make stuff.

Also in a conversation about measuring energy metrics, I'm surprised no-one has mentioned Odum's idea of energy quality and the importance of differentiating between say high intensity energy from oil and lower intensity energy from solar power.
EMergy could become an increasingly important concept for sustainabile development.

It's very easy to read too much into energy intensity numbers.

Vaclav Smil's book, "Energy at the Crossroads", devotes a lot of attention to the issues and confusions - but I also found despite all his caveats he still uses the numbers without enough care.

First of all is the purchasing-power-parity issue. What is a "dollar of GDP"? The EIA uses exchange rates, which are clearly artificially kept low in some areas (such as China) which then artificially inflates their effective energy intensity (per $ of GDP) numbers.

Second is the issue that was raised in the other comments - there are environmental differences between the nations that explain some of this even without looking at the "cultural" explanations - Canada's colder, and has greater typical travel distances than most other nations.

But third is the issue of economic structure and international trade. The more an economy is based on "services", particularly electronic services, and the less it relies on energy-intensive productive activities such as manufacturing, the lower its energy intensity will naturally be. BUT as the western economies have moved much of their manufacturing to China in recent years, there's a very large artificial element induced by international trade, that pushes China's energy intensity higher and that of western nations lower. Since oil production is a particularly energy-intensive activity, the more that has been moved out of western nations and to places like the Middle East, South America, etc., the same effect applies.

I don't know of ANY analysis that has properly taken all these elements into account. The only really useful comparison available, I believe, with our now interconnected world, is total world intensity from one year to another. Comparisons between nations, while fun, are largely faulty for these reasons.

The reason US energy use declined from 2000 was the economic recession - we still have not recovered. The recent climb in oil prices has also helped curb growth in energy use.

I think most people agree that resolving the energy problem through economic failure is not a good solution.

Ian, Japan's an interesting case because, in the past, they've had an even higher rate of oil consumption as a percentage of total energy consumption.

This chart shows that the "degree of oil dependence" for Japan peaked at 77.4 at the time of the first oil shock, and had only dropped to 71.5% at the time of the second oil shock. They responded to those two periods by bringing it down to the level of around 50% today. So, it's a lot lower than it has been in the past.

You can see that the changes from the early 70s to today have gone from less use in industrial consumption towards transport and other uses (see here) -- probably about a 1/3 reduction in industrial consumption but an almost 2 1/2 fold increase in transportation consumption.

For comparison, here are the same graphs for the United States.

Eyeballing it, it looks like you see about the same relative decrease in industrial consumption in the US, but not nearly as great an increase in transportation consumption.

Japan outsources production probably as much as the United States does, if not more so, since they face the same high labor costs and stringent environmental regulations domestically. Certainly they have shifted their production mix away from low value-added prodcuts (like metals) into higher value-added products (like silicon chips) over the years, and that would mean less energy use for production.

Joseph, Arthur, Ian,

Just wondering about the comments that "all of heavy industry" or "US outsources production" or "economy still hasn't recovered."

These were given as reasons why US appears to have gotten more energy efficient recently. I have to disagree. Reading stories in newspapers or perhaps anectdotally you would think that ALL of US manufacturing has moved offshore. But the facts do not bear this out... at least not the ones I've seen. - see facts below

First, imports are still a VERY SMALL part of the US economy. Sure, some industries have moved production to Asia or Mexico but others have continued to grow and for many it just doesn't make sense to move production offshore an have to pay extra shipping plus longer lead times etc..

Second, I think that most movement of manufacturing offshore was done in the 90's. If this is the case then why the improvement in efficiency since 2000. The latest outsourcing that we are seeing involves software,call centers,support,order entry, etc... Obviously this production would be very energy efficient and would tend to skew stats in the opposite direction. In other words, lately we have moved more energy efficient production offshore.

Third, the idea that we still haven't recovered from recession is just plain wrong. As I stated the GDP numbers are now more than 10% higher now versus 2000 and this includes the 2001 recession. GDP grew 4.4 % in 2004 and 3% in 2003. These are very strong numbers.

So I'm saying that using movement of production offshore and a poor economy as reasons for the HUGE( 10+% in GDP vs. 1+ percent in energy usage) increase in energy efficiency since 2000 just doesn't hold water.

US transportation energy usage has risen dramatically because of increased miles driven and a huge increase in SUV and truck sales. However,there have been many improvements in other
areas of energy usage. I believe that if we can increase gas mileage, we can continue to grow US economy at 3% per year AND decrease total energy consumption.

Here are a few items I dug up regarding recent industrial production.

According to Federal Reserve US Industrial Production is now at an all time high.

From the Auto Spectator regarding steel:
* Utilization: With the surge in demand, industry consolidation and price spikes, capacity utilization rates have spiked to 10-year highs for U.S. steel manufacturers. The steel manufacturing industry’s capacity utilization rate rose dramatically in 2004 to 94 percent rising from a recent low of 79 percent in 2001. Utilization rates are forecast to be near 100 percent globally by 2005. Automotive suppliers by contrast are seeing their utilization decline due in part to decreased availability, reduced quality and delayed deliveries of steel.
* Import Levels: Steel import shares have stayed essentially at or below their 10-year average levels since 2000.
From the American Steel Institute
For the year 2004, steel shipments were 112,085,000 net tons, representing a 5.8 percent increase from the 105,974,000 net tons shipped in 2003.

From the Portland Cement Association
According to the latest issue of The Monitor, the 9.282 million metric tons of portland cement consumed in November 2004 was a historical high for a November month. Masonry cement also recorded a high in November. The Monitor notes that cement usage per million dollars of real construction has increased dramatically during the past four months to 179 tons per million real dollars of construction spending.

Key statistics from The Monitor:
• Cement consumption rose by 11.4% in November over 2003 levels and year-to-date was posting a 6.5% increase.

Joe, you're covering a lot of material with your comment. I'll try to address things a little more precisely to be more clear.

My comments about heavy industry and outsourcing were with respect to long-term trends, not the immediate past few years. As economies move from low-value-added to high-value-added products and services, there is a natural decrease in per capita energy use attirbutable to that change. It simply takes way more energy to make steel and aluminum than it does silicon chips, dollar for dollar.

The context of all the comments were Jamais' original observations about comparative per capita energy use and per capita CO2 emissions. I've already stated in prior posts the different factors involved in differences in per capita energy and CO2 emissions. The issue of outsourcing came up when someone claimed that Japan still makes more domestically than the United States, and I demonstrated that they were roughly equal in their decreases in petroleum consumption for Industrial production, which seems to indicate a rough sameness in outsourcing. I'm sure I could review the literature on it beyond my own experience with the subject, but it's probably correct.

I did not suggest that the US doesn't produce anything domestically anymore.

With respect to the issue of imports, in the past 20 years they have gone from being 8% of GDP to 16% of GDP. They are now within spitting distance of what we spend on all government (from fed to local) in this country, so I think saying they are a "very small part of the US economy" is simply inaccurate.

As for the degree of movement of industrial production during specific times, it would probably be helpful to get accurate specific data before drawing conclusions based on assumptions about it.

As for the relationship between GDP changes and energy use, it's pretty obvious that economic slowdowns are going to correlate with decreased energy use. If you take a look at monthly energy consumption data and group it in to quarterly numbers, then correlate it with GDP figures, you see almost perfect correlation between the direction of GDP figures and industrial energy consumption. Residential and commercial energy consumption are more seasonal (ie, far more in the winter), and transportation energy use is correlated most directly with fuel prices. Commercial energy consumption, for example, actually went down from quarter 3 to quarter 4 of 2001, which is counter to normal the seasonal trends, reflecting the economic contraction/stagnation around the same time.

As for the last two years of economic growth being "very strong" - the average real GDP growth since 1947 has been 3.4%. So 3.0% is acutally below average, not "very strong", and 4.4% is about the 20th or 21st highest figure since 1947, so it's only slightly above normal - nothing to write home about.

As for energy use from 2001 to 2004, the growth has been different in each sector. Residential and transportation have grown by 3.5% and 3.9%, respectively. By contrast, commercial and industrial have only grown 1.4% and 2.4%, respectively. This would be the expected pattern in the short term for an economy that's driven by consumer expenditures and government spending, relative to business investment.

As for the numbers from 2000-2004, GDP grew by 10.8% and energy consumption grew by 1.8%. Again, the US has been following this pattern for quite some time - ie, delinking economic growth from growth in energy consumption - so I don't think much can be read into those numbers. The only times we don't see the energy intensity numbers go down tend to be during times of war (eg, 1967-1970, 1991, and 2002). This is also to be expected.

As for your figures about steel and cement, one would expect greater use of these materials during a building boom, which we have experienced the past few years. Capacity utilization specifically doesn't tell you much because there could have been a decrease in capacity just as easily as there could have been an increase in production - you need specific figures to make that judgment.

About yearly GDP growth figures -- if you're coming out of a recession or a low growth period, a high growth figure isn't necessarily something to be lauded, as it's merely reflecting the diminished numbers in prior periods than any supposed "strength" from growth.

Hope that clarifies some of these things. I'd like to hear more specifically what point you're trying to make with all these assertions.

First, imports are not really a "small" part of the economy - the US imported $1.76 trillion worth of goods in 2004; total GDP for 2004 was about $11 trillion, so that's about 15% of the economy (see http://www.bea.doc.gov/bea/glance.htm). More importantly, the things we import tend to be energy intensive: oil for example, which accounts for at least 10% of imports in value. Aluminum is highly energy intensive, and we import a lot from Canada and Russia. We still physically import a lot of motor vehicles from Japan, despite the transport costs.

And if you look at the China import breakdown here:
there's ten's of billions of dollars worth of imports of a huge variety of manufactured materials (total about $200 billion, perhaps more than we spend on oil). The energy costs for all that manufacturing are in China's account, not ours.

Second - you're correct that most of this is old news, though it continues to grow (with exceptions such as caused by the temporary restriction of steel imports by the Bush administration). Manufacturing did continue to drop even in unadjusted dollars from 1997 to 2002, as the economic census numbers reveal:


The reason I raised this was to point out the problem comparing energy intensity in one country with that in another. With imports at 46% of US manufacturing output, we've moved at least 31% of industrial energy use to other nations. That number is likely even larger for countries (such as in some parts of Europe) that rely much more on imports for their energy-intensive manufactured goods.

So what explains the US energy intensity decline relative to GDP since 2000? Well, it all depends how you measure. Certainly US inflation has been relatively low, so just accounting for inflation, GDP since 2000 appears to have finally climbed out of the recent recession levels, even to be 10% above where it was in 2000.

Nevertheless, by other measures, we have not recovered. Jobs are still far behind where they should be, given population increases: fewer workers means less energy use. Furthermore, if instead of discounting for inflation, you compare the dollar against other world currencies, we're actually not doing well at all.

Compared to the Canadian dollar, for instance, the US dollar was worth 20% less January 1st 2005 than January 1st 2000. Worth 25% relative to the Euro. Since Canadian and European "inflation" has nominally been about the same or more than in the US, this suggests that rather than rising 10%, US GDP measured in some mythical "real" currency may have fallen by 10% or more since 2000, despite the population increase.

The relative numbers should settle themselves after time - the above suggests we're overdue for a bout of heavy inflation (which other people have warned of also). Just as comparisons between nations are suspect, comparisons within a given nation on too short a time period are likely just as suspect. Energy intensity may be a useful tool for global averages over the long term, but it just doesn't have clear enough meaning to be useful in other circumstances.


Thanks for all your comments... the point I was trying to make is that it appeared to me that the US's energy intensity had been improving lately and I personally don't think that moving mfg offshore is the main reason for this.

I found the following link -
which show the intensity since 1973 and it shows that the improvements in energy intensity are accelerating somewhat.

My purpose for analyzing this relates back to the CO2 discussion. Most discussions that I have seen show the world energy usage growing at least 3x this century as the rest of the world catches up to the US living standard - as measured by GDP.

I want to see what it would take for this "catchup" to occur with no or little increase from current worldwide energy usage. Obviously if we could do this the CO2 problem would become much easier.

One other thing - I meant to say trade deficit instead of imports in my previous comment. The trade deficit is about 5% of the size of the US economy. Maybe that's not VERY SMALL but I would consider it to be small in the sense that it does not reflect that we are moving everything offshore. Also, a good portion of the increase in this last year was due to oil.


This page contains a single entry from the blog posted on February 26, 2005 2:23 PM.

The previous post in this blog was More About Population.

The next post in this blog is The Seas of Elysium.

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

Powered by
Movable Type 3.34