IEA counts fossil fuels threefold versus wind and solar

The IEA energy statistics underestimates the role of wind and solar in the world’s energy mix. The counting method must be changed to help policymakers and investors to understand how near the world is to a transition to an energy supply dominated by solar and wind power.

Wind and solar energy have for decades experienced exponential growth and employ millions of people, but according to the International Energy Agency (IEA) statistics they still only constitute about 2% of the world energy supply. How can this be?

The answer is well hidden in an error made in a 12-year-old Statistical Manual by the IEA and OECD. As wind and solar continue their growth, this will soon need to be changed. After correction, it will become apparent that wind and solar energy already contribute about three times more to the world’s energy supply than normally reported, and that the shift to renewable energy sources comes much sooner than many decision makers are aware of.

An illustration of the error

When a solar power plant, coal power plant and nuclear power plant all produce the same amount of electricity, one would think that energy statistics would consider these contributions to the world’s energy supply to be relatively similar. This is far from true. Both the nuclear power and coal power are counted threefold relative to the solar power plant, taking into account that about three units of coal or nuclear energy are needed to make one unit of power. A similar loss happens, of course, in a wind farm or solar park, but here the IEA statistics do not make any correction to take into account everything that has been lost. The IEA merely reports the one unit of power produced and neglects that there were also about two units of energy lost in these renewable power plants.

Consequently, the IEA ends up reporting the world’s use of fossil fuels roughly threefold versus wind and solar, and the actual role of these renewable energy sources becomes greatly underestimated. In a research paper recently submitted for publication (Sauar, 2017), it is documented that this way of counting is contrary to the basic physical energy definitions that the IEA itself states that it follows. A short, popular resume of the reasoning, however, is given below as there is currently limited literature on this topic for policy makers.

Primary energy

When the IEA compares the role of the various energy sources in the world, they normally use the term “primary energy”. The definition of this term is precise, has a clear physical meaning, and is cited by the OECD and the IEA as the basis for their statistics (See for example OECD, 2016 or UN, 1997): “Primary energy consumption refers to the direct use at the source, or supply to users without transformation, of crude energy, that is, energy that has not been subjected to any conversion or transformation process.”

Primary energy supply is, simply said, the first “raw energy” that enters an engine or power plant and then transforms into “refined energy” or “secondary energy” as, for example, electrical power or mechanical power. For gas, oil and coal, “primary energy supply” is reported correctly today as the number of tons of oil, gas or coal reflect the amount consumed during the year. Whether this is used efficiently or not, is not counted. When the oil is refined and later enters a car engine, for example, most of the energy is lost, and only about 20% of the primary energy excerpted from the oil field is converted into refined, mechanical energy that drives the car forward.

For solar power plants and wind turbines, the primary energy is the solar energy that is radiating or the wind energy blowing into the power plant, while the refined energy is the electricity that is being produced. As in the combustion engine, much of the primary energy is lost, but about the same amount is turned into the desired electricity (or mechanical power). So why are these two situations not treated equally and according to the definition of primary energy? Most likely part of the reason is mere simplicity. Companies producing oil or coal count the amount that they extract in tons. Companies producing electricity from solar and wind energy count the amount of electricity they produce.

Anyway, the OECD and IEA statisticians 12 years ago selected a model (IEA/OECD, 2005) to count the consumption of a primary energy source that generates electricity based on how much heat was produced in an interim process step, even in cases where this is directly contrary to the physical definition of primary energy. When wind and solar were discussed at that time, these were very small sources of energy so the discussion was possibly concluded without too much attention.

In later years, this error has simply been repeated. This means in practice that the primary supply of solar and wind energy is underreported with factors 3-4 both today and in all the scenarios for the future.

Giampietro and Mayumi (2009)’s figure below is a good illustration of how a primary energy source moves from the left side through various conversions before it is actually taken into use. In this figure, it is easy to show where the IEA and most energy statisticians repeat their historical mistake, and since the losses in and up to conversion #1 typically are in the 60-80% range, this of course makes a major difference.

Conversion of primary energy sources as previously illustrated by Giampietro and Mayumi (2009). The blue arrows are added in this paper showing where IEA counts the different energy sources.

The physics problems with the IEA counting method

As can be read in the IEA and OECD Statistical Manual (2005), the IEA currently regards the primary energy form as the first energy form downstream in the production process for which multiple energy uses are practical. However, as can be seen in the figure above, this means that most energy sources are counted as primary energy, while some others (hydro, wind and solar) are counted as secondary energy. This is hence not a counting method which will count the well-defined term primary energy. It will probably instead count something that could be named “tradable energy”.

Even though the IEA is of course free to define its terms, the organization will need to have a definition of primary energy form which is consistent with the term primary energy. Secondly, the term primary energy is a well-defined term in the general scientific community, and is not generally open for changes outside the original term primary energy.

A second problem is that the current method counts a primary energy source 100% according to its primary energy content only if it goes through an interim heat production process. This is particularly problematic since the interim production of heat is frequently not the most energy efficient route for converting a primary energy source to secondary energy. Nearly 50% of the thermodynamic potential to perform energy work (the potential electricity) from a chemical energy source (like oil and gas) is immediately lost when converting it to heat. Oil and natural gas can in principle be converted to nearly 100% electricity in a perfect fuel cell. After these energy sources are converted to interim heat, however, they have lost nearly 50% of their potential to make electricity due to fundamental thermodynamic principles.

Appropriate counting of wind and solar as primary energy sources

For nuclear energy and biomass based power plants, it is also easier to measure the amount of electricity produced than to count the exact amount of nuclear energy that was consumed, or to weigh the exact amount of biomass consumed. The IEA has therefore chosen to only measure the electricity produced for these two energy sources, and thereafter multiply it with 3.0 for biomass and 3.03 for nuclear energy.

For wind and solar it is very easy to derive similar factors. In the longer research paper recently submitted for publication, it is showed that electricity produced from wind turbines probably would need a multiplication factor in the range of 2.2–2.5 in order to take the conversion losses into account, as this is a very energy efficient technology with energy efficiencies frequently exceeding 40%.

Solar panels, on the other hand, are lower in efficiency, but will soon be around 20% on average, and would need a factor in the range of 5.0 in order to account for the losses. This last factor may appear high, but the losses in the conversion of solar energy to secondary energy are actually fairly similar to the losses of oil based energy prior to and inside a combustion engine. For a 60/40 mix of wind and solar, the average accounting factor would be around 3.3, which is very much in line with how biomass and nuclear is treated.

Decisions are based on perception

Voters, industry leaders and politicians do not make their decisions based on objective facts, but on how they perceive and understand them. That is why perception matters. If a marathon runner knows he is about to win, he will put in more effort. If an industrial company knows that the market will prefer what they produce over the next 25 years, then it is easier to invest. Politics obviously also is an “art of the possible”. Hence, it matters very much what politicians believe is possible.

A relevant and contemporary example can be found in today’s political discussion in Norway, where the oil industry is arguing for opening more areas for oil exploration. One of their main arguments (latest used on VGTV August 15, 2017 by leader of the Norwegian Oil and Gas Association, Karl Eirik Schjøtt-Pedersen) is that the world needs more energy and that the new renewables are not mature yet to step up as they account for only 2% of the energy supply. A correct reporting by the IEA would clearly change this perspective and weaken or remove this argument. That again is likely to impact what decisions voters and politicians will make.

One can, of course, argue that since wind and solar radiation are eternal sources of energy for all practical purposes, it makes perhaps more sense to report these as secondary energy, i.e. electricity. However, in that case one would need to report also all the other energy sources as secondary energy in order to show what relative role the different energy sources play in our societies.

Wind and solar are soon to dominate the world energy supply

The figure below shows a comparison of the world’s use of primary energy in one of the three IEA policy scenarios for year 2040. This is the so-called 450 scenario, where the amount of carbon dioxide in the atmosphere is constrained at 450 ppm in order to limit climate change.

To the left is the energy supply reported by the IEA in their annual World Energy Outlook 2016 for this scenario. To the right is the same graph, but showing the primary energy consumption as the metric also for wind and solar energy. The difference is more than significant, with wind and solar energy now clearly taking the role as the largest source of primary energy. In the graph to the right a multiplication factor of 2.9 has been used based on the mix of wind, solar PV, Solar CSP and some geothermal power that the IEA has in this scenario for 2040. (There are also some other renewable energy sources reported together with wind and solar by the IEA, which may contribute to modifying this factor slightly when done even more precisely.)

The importance of counting energy: IEA 450 scenario (left) and corrected version (right) give very different impressions of the future role of wind and solar.

It is worth noting in the graph above that on the left side nuclear power appears to play the same role as the combination of wind and solar in 2040. However, this is a pure artefact of today’s IEA counting since the amount of power produced from wind and solar is foreseen to be a lot larger than that from nuclear in 2040. This is hidden in today’s reporting, but becomes very clear when wind and solar energy is correctly accounted for.

Hydro power is worth giving a special comment. This energy source is in its primary energy state a kinetic energy (or mechanical energy) that can be converted to electricity with nearly 100% efficiency. Hence, for hydro power the primary and secondary energy amount are nearly identical.

If you are an industry leader considering investing in fossil fuels and look at the graph to the left, you would believe that the competition from wind and solar energy would hardly hit you until after 2040. If you look at the graph to the right, however, it is clear that this competition will be severe well ahead of 2040.

For voters and politicians considering whether or not they should stimulate wind and solar energy it is likewise very discouraging to see that even after 25 years of stimulation for solar and wind energy, the fossil energy sources are still dominating. Such perceptions can easily lead to hopelessness and passivity. A correct illustration of the actual importance of wind and solar energy relative to the others is hence of high and practical value. It matters.

Today’s counting method leads politicians, industrialists and voters to believe that a shift from fossil fuels to renewables is much farther away than it actually is.

Oil vs renewables for transport

In the introduction, we looked at energy used for power production. However, a lot of energy, oil in particular, is used for transportation purposes. There is currently a beginning mass supply of electric vehicles combined with strong growth in wind and solar energy. This development is particularly relevant because electric vehicles are a very good way to store the intermittent energy from wind and solar power plants. It is also particularly relevant because there are normally significant losses when converting any energy source to electricity or usable mechanical energy. For wind and solar energy, this conversion loss occurs at the power plant. For the oil based energy, this conversion loss occurs in the combustion engine.

Let us therefore compare the two value chains conventional-car-based-on-gasoline and electric-car-based-on-a-mix-of-wind-and-solar energy. Ideally, this comparison should be done based on global averages, but these are not always readily available. We have therefore below only presented reasonably representative numbers. A more thorough global study would obviously become more precise, but the result would probably not be radically different. The consumption data for an average gasoline car are for example taken from the US EPA (Thomas, 2017).

VehicleFuelPrimary energy
per 15 000 km
(scientific definition)
Primary energy
per 15 000 km (Today’s IEA
Gasoline carGasoline17 MWh1
(EPA, 2017 and Keesom, 2012)
17 MWh
Electric carWind/Solar
15 MWh24 MWh3
Footnotes:1. Calculated from an average 25 MPG reported as the average consumption for 2017 car by EPA combined with an average 80% extraction efficiency reported by Keesom et al. (2012) and using 36,6 kWh per gallon. 2. Calculated from a typical 25 kWh per 100 km, assuming a typical 10% transmission losses for electricity, and assuming that the electricity was produced from renewables with a 28% conversion efficiency based on a 50/50 mix of solar and wind power. 3. Calculated from a typical 25 kWh per 100 km and assuming a typical 10% transmission loss for electricity, but assuming (erroneously) that the primary energy of a wind or solar power plant equals the electricity produced.

As can be seen above, the two value chains are actually very comparable in primary energy consumption per kilometre if the actual definition of primary energy is used with a gasoline car consuming 17 MWh per 15 000 km while the electric car on solar power consumes about 15 MWh. Taking into account that the average car in the US may be larger than the average electrical car, the small benefit of the electric car will probably even disappear.

Today’s reporting by the IEA, however, reports 75% less primary energy usage for the renewable energy value chain and thereby again tends to completely overstate the role of the fossil fuels in our economy.

The “primary energy error” of current energy statistics thereby has the slightly absurd impact that whenever a car converts from fossil fuels to solar and wind power, the world energy primary consumption appears to fall by 75% even though that is not happening! In many of the toughest climate policy scenarios, they therefore show that the world primary energy consumption is squeezed downwards over the next 20 years, and the reader gets the understanding that this is a tough energy diet to take. But as can be seen above this is simply not true, but an artefact of the current statistical counting method.

Two different solar technologies illustrate well today’s counting inconsistency

Solar electricity is today generated through two different technologies; Photovoltaic (PV) systems that convert solar energy directly to electricity (with a low temperature heat loss) and Concentrated Solar Power (CSP) which converts solar energy first to high temperature heat and thereafter converts the heat to electricity in a more regular thermal power plant with a similar low temperature heat loss.

For the Solar PV technology the IEA counts only the electricity produced, as has been thoroughly explained above. For the Solar CSP technology, however, the IEA counts the interim high temperature heat production and thereby reports it with a 3 times higher primary energy production than solar PV – even when they produce exactly the same amount of electricity from nearly the same amount of solar energy. And in both cases, the conversion loss even ends up being low temperature heat. For a reader of the world’s energy supply scenarios, CSP hence always appears to have a fairly significant role, but that is simply because the power production from a CSP plant is multiplied with factor 3 and PV is not. The situation can be illustrated in the figure below:

IEA’s accounting of two different solar energy technologies.

Both today and in the foreseeable future, PV will dominate between these two – partially due to cost and partially because CSP requires near-desert weather in order to deliver low cost energy production.

Are there good reasons to maintain the IEAs current counting practice?

Scientific consistency
First and foremost, it must be a clear requirement to international strong-reputed organizations like the IEA that their counting methods are in accordance with the physical entities they aim to count. The IEA may of course argue that they follow the thermal based counting method that they describe in the OECD and IEA Statistical Manual, but this counting method does not count primary energy when it comes to solar and wind power. It counts only the secondary energy made. The IEA should of course therefore report their current tables as a “Mix of primary and secondary energy” or as “Tradable energy produced” in order to have their science consistent. Merely the case for having international organizations use scientific terminology precisely should be sufficient to change today’s practice – either by changing the counting method or by using different terminology so that there at least is consistency between numbers and terms.

Most relevant perspective
Another important perspective is, of course, which numbers that give the most relevant perspective on the contribution of the different sources of energy to the world’s energy supply today and in the future. A significant question could be raised if there existed a large energy resource with only 5% energy efficiency as this would quickly appear to be much more relevant than it is. Fortunately, that is not the case.

Based on the table above for the transport sector, it should be clear that using the term primary energy correctly also for wind and solar, will give a much more relevant understanding than today’s inconsistent terminology. For the power sector, it is similarly clear. Solar modules will shortly have an efficiency of about 20%, while wind turbines are frequently above 40%. This is not very different from nuclear, biomass and coal power plants, especially when also taking the transport and refinery losses into account prior to the power plants.

BP (2016) has realized this problem and therefore recently started to multiply the contributions from wind and solar by a factor of 2.8 in order to better show the actual contribution from wind and solar energy to the world’s energy supply in their scenarios. This factor was chosen in order to take into account the average losses in a fossil fuel power plant, and thereby estimate what amount of fossil fuels that wind and solar replace. In order to account also for all the extraction and transport losses in order to get the oil, coal and gas to the power plant, the conversion factor should probably increase to about 3.1–3.4.

As shown earlier, however, if one looks at the actual losses in converting wind and solar energy to electricity, the proper conversion factor would probably be about 2.3 for wind and 5.0 for solar power. The average of this is about 3.3, so both perspectives actually lead to roughly the same conclusion.

In this discussion, it is also relevant to compare with nuclear power and biomass power, where the IEA has chosen to count the primary energy as 3.03 and 3.0 times the electricity produced. Introducing similar factors for solar (~5.0) and wind energy (~2.3) would thereby immediately establish a consistent counting method also with respect to these energy sources.


This article shows that today’s counting method by the IEA and OECD energy statistics is not scientifically consistent. It also shows that if the methods are modified so that solar and wind are also represented by their contributions as primary energy sources, the audience will get a better understanding of the relative contribution of wind and solar energy in today’s energy mix.

Finally, it is clear that making a change in today’s counting method is important and urgent. Companies, governments and voters make their decisions based on how they perceive the world. Understanding how near the world is to a transition to an energy supply dominated by solar and wind power is likely to impact investment decisions and political decisions in many parts of the world.

List of references:

BP Energy Outlook, (accessed 28.05.17), 2016.

Giampietro, M. and Mayumi, K., The Biofuel Delusion: The Fallacy of Large Scale Agro-Biofuels Production, Earthscan, 2009.

IEA, World Energy Outlook 2016, Paris, 2016.

IEA/OECD, Energy Statistics Manual, OECD Publishing, Paris, 2004.

Keesom, B., Blieszner, J. and Unnasch, S., EU Pathway Study: Life Cycle Assessment of Crude Oils in a European Context, Jacobs Consultancy, 2012.

OECD, OECD Glossary,, (accessed 28.05.17), 2016

Sauar, E.: submitted for publication, 2017

Thomas, J., Drive Cycle Powertrain Efficiencies and Trends Derived from EPA Vehicle Dynamometer Results, SAE Int. J. Passeng. Cars – Mech. Syst.7(4) : 1374-1384, doi:10.4271/2014-01-2562, 2014.

UN, Glossary of Environment Statistics, Studies in Methods, Series F, No. 67, New York, 1997.

VGTV,!/video/145098/oljetopp-i-strupen-paa-miljoebevegelsen, August 15, 2017