For decades economists have tried to link energy use and GDP. However, this has grown more tortured over time, as GDP gets increasingly divorced from the material inputs that consume large quantities of energy. As developed countries shift more of their efforts into services and out of industrial production, our GDP grows quickly, as services ultimately give more bang for the buck, and this flatters our figures regarding energy intensity.
This leads to tortured projections that are increasingly less useful.
Other analysis focuses on supply constraints and costs. The Department of Energy’s Energy Information Administration is really keen on showing how much fossil fuels are going to cost and make that a key part of their models of future consumption. But in country after country we see that, once a certain level of per capita income (and not GDP) is reached, energy is not truly constrained by supply–people will pay whatever it costs. Modelling costs and availability is really a good idea for countries like India. It’s not nearly as useful for countries like the U.S. This is one of the reasons why the DOE’s EIA has to keep adjusting figures.
However, energy consumption is far more tightly linked to population. And every analysis pays lip service to that fact. The EIA Annual Energy Outlook 2012 says, “Population is a key determinant of energy consumption through its influence on demand for travel, housing, consumer goods and services.” And then they promptly move on to other topics.
The fact is that we consume one quad of energy for every 3 million inhabitants in this country. If we get 3 million more immigrants or children born tomorrow, we will consume one more quad than we otherwise would have.
Population is the principal component of energy consumption.
Our population is growing.
You will have to make a very good case to convince me that the easiest way to estimate future energy consumption in America is anything other than to add one quad to our total for every cohort of 3 million new humans in the U.S. of A. By way of checking, our energy consumption was more than 1 quad per 3 million people in 1990, and also in 1980–so I don’t think I’m exaggerating for effect here.
The U.S. Census Bureau has published estimated populations through 2100. They took four guesses–the series was published in 2000 and their highest growth series is bang on accurate. Most estimates use what they called their Middle Series, but their Middle Series predicted a 2010 population of less than 300 million. If it’s that far out after just a decade, I don’t really want to use it for a century. What they called their High Series was much better, so I’m going with it. (But I’ll show a couple of figures later with the Middle Series.)
Our current population is 311 million. We consumed 98 quads in 2010. So here are the Census Bureau’s High Series projections at 10 year intervals, with my highly sophisticated algorhythm for calculating additional energy consumption (add 1 quad for every 3 million people):
2020 354,642,000 people 113 quads
2030 409,604,000 people 130 quads
2035 441,648,000 141 quads
2040 475,949,000 people 153 quads
2050 552,757,000 people 179 quads
2060 642,752,000 people 209 quads
2070 749,257,000 people 244 quads
2075 809,243,000 people 264 quads
2080 873,794,000 people 286 quads
2090 1,017,344,000 people 333 quads
2100 1,182,390,000 people 388 quads
Sobering totals. Obviously, the DOE used the Census Bureau’s Middle Series for their projections, so let’s show a couple of years using that list. But for the moment, let’s hang on to my little conceit that calculating future energy consumption can be simple arithmetic, rather than convoluted conjecturing.
2035 353,749,000 people 112 quads
2050 403,687,000 people 129 quads
2075 480,504,000 people 155 quads
2100 570,904,000 people 185 quads
Now remember that the EIA only projects energy consumption through 2035. Their projection for that year was 108 quads, revised down from 114 quads because they felt that energy efficiency and (???) an improving economy (???) would lower energy consumption. My back of the envelope calculations show 112 quads for 2035 using their choice of the population estimates, and 141 quads if the High Series continues to be more accurate.
My totals amount to 0.61% CAGR using the Middle Series and 1.47% using the High Series. Remember also that last year’s growth in energy consumption was 3.9%.
Population growth using the High Series from the Census Bureau is 1.49% CAGR. Using the Middle Series shows population growth at 0.68%.
So I would say we have a range of possible energy consumption figures that vary between 112 and 141 quads at 2035, between 129 and 179 quads at 2050 and between 185 and 388 quads at 2100.
This would seem to put into question the EIA’s projections of future energy consumption for the U.S. If the High Series is accurate, and if my crude metric is closer to reality than the EIA’s supply-dominated but sophisticated model, energy growth in the U.S. will be 1.54% per year, rather than the 0.39% the EIA forecasts.
Sadly, given what we have seen regarding Mexico and Turkey, this would almost force us to conclude that the EIA was as overly optimistic about energy consumption in the OECD as it was for developing countries.
Now, honestly, it is my hope that the light-hearted tone of this post is enough to convince readers that I don’t really intend to suggest that a mere calculation of population growth is an adequate substitution for serious thought about energy consumption. There is more to it than just simple arithmetic. But serious thinking is more than just additional calculations.
Population growth generates momentum towards a certain level of energy consumption. Efforts to lower than consumption through energy efficiency and innovation will fight strong headwinds created by this population growth. But just as my argument is not totally serious, neither are arguments that energy efficiency alone will dramatically lower consumption. The difference is that I acknowledge the incompleteness of my arguments…
On of the main problems with linking GDP and energy consumption is that because of globalisation, it is hard to measure “true” consumption. This is because items like cars, concrete and steel are made overseas. The goods are imported so their manufacture doesn’t show up in the importer countries statistics. I don’t think anyone has a good idea of all the inputs into even a simple item, let alone something complex like say a hybrid car or a laptop.
I have no data to back it up, but I suspect when all the inputs were properly apportioned, GDP to Energy over time would show a lot better correlation than say a hockey stick 😉
You may have a point–hockey sticks do seem to be out of favor these days. And GDP/energy consumption may be masked by imports/exports. And I am among the many whom you describe as not having a good idea of all the inputs even to simple items. That said, I do think the shift to services is making GDP increasingly less useful as a metric.
Yes.Ii think your projections are a lot more realistic than those of EIA. All the energy saving and high efficiency devices haven’t lowered energy demand, they have just slowed the growth rate.
We keep on being promised a green revolution or the like, but there has been no data or even examples to back that up.
The USA is a hard economy to study for energy because of its size, diversity and porous borders. it is a lot better using island nations like Iceland or New Zealand. They also generate a lot of their electricity from renewables so data is more complete. In NZ, the electricity consumption has grown at an almost steady pace of 800GWh pa since 1959! The percentage growth each year is declining but the trend is still up. Household consumption continues to climb even with government sponsored efficiency programmes. Higher efficiency has translated to warmer rooms or more boy’s toys – personal behaviour trumps idealistic principles.
Pingback: Global Energy Consumption in 2030: Preliminary numbers and discussion of confidence levels | 3000 Quads