Skip to content

Michael Kelly: A Challenge For Renewable Energies

Prof Michael J Kelly, Department of Engineering, University of Cambridge, Cambridge CB3 0FA, UK

This paper examines whether renewable energy can match the economic productivity of the current, largely fossil energy supply and the consequences if it cannot.


In recent years the energy sector has accounted for about 9 per cent of global GDP,[1] with the implication that the return on energy investment in the world economy is, approximately and as an average, about 11:1. In other words, the energy produced with 9 per cent of the world’s economic activity is enough to power the whole world economy, from profligate energy consumption in the US through to fuel poverty in sub-Saharan Africa.

The bulk of global energy has been and still is provided by fossil fuels, which have stood at a remarkably constant 87 per cent of supply since 1990.[2] In fact, the narrowly defined energy returns on investment for individual fossil fuel technologies are far higher than the global figure, being of the order of 50:1 for coal and gas plants and 70:1 for nuclear reactors. Only when combined with the low values for traditional biomass and all the societal externalities does the effective overall value fall to the 11:1 cited above.[3]

This paper examines whether renewable energy can match the economic productivity of the current, largely fossil, energy supply and the consequences if it cannot.

Scene setting: past and present trends

The number of people in the global (lower)[4] middle class[5] has doubled in the last 20 years: from 1.5 billion in 1995 to 3 billion out of 7 billion in 2015.[6] Most people would hope that this trend will continue, lifting the whole of the global population out of grinding poverty by the time the world population peaks at over 10 billion at or beyond about 2065–75.[7]

During the same period, 1995-2015, total world demand for energy has increased by 40 per cent, because a middle-class person uses approximately 3.5 times as much energy per day as a person in an urban slum or a rural hovel. The burgeoning middle class in China alone can explain 40 per cent of this growth.

The forward projection is that by 2035 there will be 5 billion middle-class people out of a world population of 8 billion, and that demand for energy will rise by another 40 per cent. However, renewables are expected to supply only a small part of that additional demand; well over 80 per cent will still come from fossil fuels.

On top of that advance of the poor into the middle class, those currently in the middle class will become rich, with access to advanced mobility, communications, healthcare, diet, education, and so on, as described above. The working hypothesis here is that poorer people will aspire eventually to enjoy a European though not a North American standard of living in that distant future. Europeans use about seven times as much energy per person per day[8] as their predecessors did in 1800, an amount comparable with energy consumption of the poor today. I will therefore assume that a 7:1 ratio applies in the transition from poor to rich, rather than the figure of 3.5 for transition from poor to middle class. Rich people use twice as much energy per person per day as the (lower) middle class.

There are no World Bank estimates of the future growth of the global rich, and this growth rate could be a parameter for a fuller analysis. However, for the purposes of this initial discussion of renewable energy, it is assumed that there will be no growth in the number of rich people, which is obviously a conservative estimate. To the extent that the rich population does grow, the conclusions drawn here will need to be strengthened further, as described below.

Renewable energy: solar and wind

The bulk of the present generation of so-called “new” renewables is provided by wind and solar photovoltaic technologies. While the latter is only 40 years old as a technology, it is worth noting that wind power is, strictly speaking, 4,000 years old.

There has been one serious and comprehensive study of the energy return on investment of the solar photovoltaic industry, using data from Spain during the period 2009-11.[9] Because of legislation concerning access to the funding of this revolution, all the relevant data on money for energy returned on money invested in the solar farms is in the public domain and is remarkably complete.

One takes the revenue for the first three years and extrapolates out at full capacity (although as a matter of fact the amount of energy generated by solar farms declines at 2–3 per cent a year) over the 25-year design life of a solar farm to get the energy money return. The total cost of establishing the solar farms includes many factors, including land rents for twenty-five years, construction of roads for installation and maintenance access, the manufacture of panels and their stands and their transport to the site, their installation and connection to the grid, maintenance, legal fees for permissions, financial services, and so on. These add up to about 40 per cent of the above revenue (30 per cent of the cost is the panel manufacture, 30 per cent the rents and 40 per cent all the other factors). Thus we can calculate a 2.5 ratio for energy money return on energy money invested, which is less than the rate for the prevailing global economy.

Suppose over the intervening ten years the rents and other costs were frozen, but the panels became free; the ratio of energy revenue on energy money invested would still be only 3.7. If the solar panels were to represent a direct replacement for fossil fuel energy production, they would have to produce 3–4 times more energy per panel than they do at present, and at no extra cost, in order to maintain the ratio of 11:1 for the global economy.   Note that others try to assign the costs in different ways and get higher values of this ratio, but I am using a cost/revenue analysis of a well-isolated project for which all the numbers are available[10].

The situation of wind power is no better. Whereas 300 tonnes of steel in a combined cycle gas turbine can use natural gas to produce a generator with a capacity of 600 MW (or 2 kW per kg of steel), the same kilogram of steel in the nacelle of a wind turbine contributes 2 Watts of capacity, and once the concrete in the plinth and foundation of a wind turbine is included the ratio deteriorates still further. While the energy return ratio for wind is higher than solar photovoltaics, it is at most double that figure, and therefore still rather smaller than the 11:1 value that characterises the world economy today.

Moreover, once we include batteries as a form of ensuring despatchable electricity, electricity on demand, the return on investment declines further. Batteries, or any other form of energy storage, simply mean that more cost is expended for no more total energy delivered, therefore the cost to consumers must rise.


Suppose in the future much improved renewable technologies expand to meet 50 per cent of the world energy demand, while operating at a value of energy money returned on energy money invested of 5.5 (i.e. half that which applies today in the global economy), while the other 50 per cent of the energy supply is as it is today. In this scenario, the 11:1 ratio would decline to 8:1 in round figures, and we would have to reduce overall global economic activity by about 40 per cent. For the first 10 per cent cut that could be achieved by eliminating or severely curtailing high culture, tertiary education, advanced medicine, international travel and the internet. I will leave it to readers to make their own choices as to how to extend this list to deliver the full 40 per cent reduction. Clearly, this is not an attractive proposition that would win democratic support.

In fact the best estimate from BP and Exxon Mobil is that renewables as we know them will barely get to produce 10% of the world primary energy demand by 2040, and so the short list of draconian changes implied above would be sufficient, though these are no more likely to be acceptable than the deeper reductions.[11]

It would seem, therefore, that to the extent that more people will want to become rich, the world will need more fossil and nuclear fuels, not less. There is one way to try to square this circle and that is to redesign the world to be more energy efficient, and for those who are rich and middle class to halve their per capita daily energy consumption by changing their lifestyle, but, again, this an unlikely prospect.[12]

Suppose instead we try to envision today’s global economy being run on the basis that the energy sector is 50 per cent of the total economy, with an implied energy return on investment of 2:1, as would be the case if current renewables become pervasive, as per the recent study much discussed study by Jacobson.[13] What would the economy look like? To support the rest of the current economy, a further 90 per cent increase in population to service the expanded energy industry, and an increase of 90 per cent in the world GDP would be required. To first order, both of these changes are impossible. Furthermore, the second order interaction effects would be enormous and serve to shrink the global economy rather further than described above to meet the population constraint.

Finally, if all the approximately 10.3 billion people likely to be on earth in 2065 were to live European lifestyles, energy demand would be over four times that which we use today, less the considerable saving expected from energy efficiency and conservation.

Thanks to John Constable and Andrew Montford for refining the original text.


[1] A Primer on Energy and the Economy: Energy’s Large Share of the Economy Requires Caution in Determining Policies That Affect It

[2] Most of the data cited here is collected in: M J Kelly, “Lessons from technology development for energy and Sustainability”, MRS Energy & Sustainability,: A Review Journal, Vol. 3 pp 13 © Materials Research Society, 2016, doi:10.1557/mre.2016.3

[3] Kelly, opus cit.

[4] Author’s parenthesis.

[5] Those living in a dwelling with electricity and running water, but without advanced mobility, communications, healthcare, diet, and education, and other things that are the province of the global rich.

[6] Kelly, opus cit

[7] Kelly, opus cit

[8] Kelly, opus cit

[9] Pedro A Prieto and Charles A S Hall: Spain’s Photovoltaic Revolution: The Energy Return on Investment, Springer 2013, or the analysis on The net EROI for solar PV: a case study for Spain

[10] R. H. E. M. Koppelaar, (2016). Solar-PV energy payback and net energy: Meta-assessment of study quality, reproducibility, and results harmonization. Renewable and Sustainable Energy Reviews 72, 1241–1255

[11] Kelly, opus cit.

[12] B. Bajzelj, J. M. Allwood, and J.M. Cullen, ‘Designing climate change mitigation plans that add up’, Environ. Sci. Technol. 47, 8062 – 8069 (2013 ).

[13] Jacobson MZ, Delucchi MA, Cameron MA, Frew BA (2015) ‘Low-Cost Solution to the Grid Reliability Problem with 100% Penetration of Intermittent Wind, Water, and Solar for All Purposes. Proceedings of the National Academy of Sciences of the United States of America 112(49):15060–15065. For potential problems with this projection, see Christopher T. M. Clack et al, 2017 ‘Evaluation of a proposal for reliable low-cost grid power with 100% wind, water, and solar’ Proceedings of the National Academy of Sciences of the United States of America, published ahead of print: doi/10.1073/pnas.1610381114