Coronavirus pandemic: daily updated research and data.
Read more


First published in 2015; most recent substantial revision in July 2018. This article previously covered aspects of energy access, including access to electricity and per capita consumption; you now find this material in our entry on Energy Access.

Access to energy is a key pillar for human wellbeing, economic development and poverty alleviation. Ensuring everyone has sufficient access is an ongoing and pressing challenge for global development.

However, our energy systems also have important environmental impacts. Historical and current energy systems are dominated by fossil fuels (coal, oil and gas) which produce carbon dioxide (CO2) and other greenhouse gases– the fundamental driver of global climate change. If we are to meet our global climate targets and avoid dangerous climate change, the world needs a significant and concerted transition in its energy sources.

Balancing the challenge between development and environment therefore provides us with an ultimate goal of ensuring everyone has access to enough sustainable energy to maintain a high standard of living.

In this entry we attempt to cover the fundamental pillars we need to understand global and regional energy systems: their evolution through time in terms of consumption, relative sources, and trade; progress in global energy access and our transition towards low-carbon sources; and crucially the main development, economic and health drivers behind the energy choices we make. It is intended to provide a fundamental background to the macro-trends in our historical and current energy systems, with key learnings on how we can use this understanding to shape pathways towards a sustainable future.

How much energy does the world consume?

Let’s first take a look at how global energy production- both in terms of quantity and source- have changed over the long-term. In the visualisation we have plotted global energy consumption from 1800 through to 2018. Note that you can use the absolute/relative toggle on the chart to view these in absolute numbers or as the percentage of the global total.

Energy production by region

How are total levels of consumption distributed across the world’s regions? In the chart we see primary energy consumption from 1965-2015 aggregated by continental regions. Note that this dataset only includes commercially-traded fuels (coal, oil and gas), nuclear, and modern renewables. This means traditional biofuels are not included; as a result, figures are likely to be a small underestimate for regions (predominantly Africa and developing Asia) where populations still strongly rely on traditional biomass as a primary fuel source.

In 1965 the bulk of total energy was consumed North America, Europe and Eurasia- collectively, they accounted for more than 80 percent of global energy consumption. Although energy consumption has increased in these regions since the 1960s, their relative share of the total has declined significantly. Consumption across the rest of the world has been increasing, most dramatically in the Asia Pacific where the total consumption increased more than 12-fold over this period.

As a result, in 2015 Asia Pacific was by far the largest regional consumer with 42 percent- this was about the same as North America, Europe and Eurasia combined (at 43 percent). The Middle East, Latin America and Africa account for around seven, five and three percent, respectively.

Energy consumption by source

In the visualizations we compare the breakdown of energy consumption by source.

This can be shown in two ways – as ‘primary’ energy consumption, and as energy consumption ‘corrected’ for inefficiencies in fossil fuel conversions.

Primary energy consumption is often called the ‘direct method’ as it shows energy statistics exactly in their raw form: how much coal, oil and gas energy are consumed as inputs to the energy system.

But, this approach does not account for the inefficiencies that fossil fuels incur when converted into final energy. We therefore show values corrected by what is termed the ‘substitution method’ – this gives a better approximation of final energy demand, and is often viewed as a more appropriate way to compare the shares of different energy sources.

Long-run energy transitions in countries 

While most people associate the advent of energy with the uptake of coal, it’s important to understand what modern fuels have replaced by taking a long-term perspective on the evolution of human energy systems. In the chart we see long-term trends in energy transitions in Italy; this figure has been developed based on data from Gales et al. (2007).1

Similar data across a range of countries in Europe and the Americas has been made available at the Energy History project at the Joint Center for History and Economics, Harvard University and University of Cambridge you can explore these trends using the “change country” function in the chart.

These trends provide an additional energy dimension: human and animal power. The inclusion of muscle, food for labour and animal feed reminds us of the important earlier transition in these economies from human and animal labour to industrialised energy production. In high-income countries, the uptake of fossil fuels- and later, the integration of renewable and nuclear technologies- has effectively eliminated the use of human or animal labour. In some low-to-middle income nations, the contribution of a human labour force (especially in agricultural and manufacturing sectors) is still significant, but continuing to progress through the composition shifts we see in the figures.

Per capita energy consumption

Here we see trends in per capita energy use from 1960-2014; this is inclusive of all dimensions of energy (electricity plus transport and heating), not exclusively electricity (with energy normalised kilowatt-hour equivalents per year). There are several important points to note. Firstly, global average per capita energy consumption has been consistently increasing; between 1970-2014, average consumption has increased by approximately 45 percent.

This growth in per capita energy consumption does, however, vary significantly between countries and regions. Most of the growth in per capita energy consumption over the last few decades has been driven by increased consumption in transitioning middle-income (and to a lesser extent, low income countries). In the chart we see a significant increase in consumption in transitioning BRICS economies (China, India and Brazil in particular); China’s per capita use has grown by nearly 250 percent since 2000; India by more than 50 percent; and Brazil by 38 percent.

Whilst global energy growth is growing from developing economies, the trend for many high-income nations is a notable decline. As we see in exemplar trends from the UK and US, the growth we are currently seeing in transitioning economies ended for many high-income nations by over the 1970-80s period. Both the US and UK peaked in terms of per capita energy consumption in the 1970s, plateauing for several decades until the early 2000s. Since then, we see a reduction in consumption; since 2000, UK usage has decreased by 20-25 percent.

Nonetheless, despite this decline in high-income countries, large global inequalities still exist. The average US citizen still consumes more than ten times the energy of the average Indian, 4-5 times that of a Brazilian, and three times more than China. The gulf between these and very low-income nations is even greater- a number of low-income nations consume less than 500 kilograms of oil equivalent per person.

Energy intensity of economies

If we want to continue growing economically, increasing prosperity, and working towards poverty elimination (which most countries and individuals do) whilst efficiently managing energy resources (and reducing greenhouse gas emissions), ‘energy intensity’ becomes an important metric for tracking progress. Energy intensity measures the quantity of energy needed to produce one unit of gross domestic product (GDP) growth. It’s typically measured in kilowatt-hours of energy needed to produce one dollar of growth (kWh per dollar). It is essentially a measure of the energy efficiency of economies; we want to achieve economic growth with as low an energy input as possible.

In the chart we show how the energy intensity of economies have changed since 1990 (measured in kWh per 2011 international-$). Here, we see a distinct downward trend- at the global level, as well as across all income-level brackets. Note that you can view trends for individual countries on the interactive chart, and get a global overview using the ‘map’ tab.

In 1990, as a global average, it took 2.1 kWh of energy to produce one international dollar of economic output; in 2014 this had declined to 1.5kWh. This represents a 30 percent reduction. Efficiency gains have been seen across all income-levels. High-income economies typically have the lowest energy-intensity (i.e. they are more energy efficient per unit of economic output), and a large efficiency gap exists between lowest-income nations and the rest of the world. The relative energy intensity of economies is strongly linked to their composition, and more specifically the share of services versus industry and manufacturing output. The links between energy intensity and economy composition are discussed later in this entry.

Are we making progress on decarbonization?

If we want to reduce our global greenhouse gas emissions, the world has to transition from an energy system dominated by fossil fuels to a low-carbon one (this is what most countries have set long-term targets to achieve within the Paris climate agreement).2

With the exception of carbon capture and storage (CCS) technology (described later in the entry), we have two options to achieve this: renewable technologies (including bioenergy, hydropower, solar, wind, geothermal, and marine energy) and nuclear energy. Both of these options produce very low CO2 emissions per unit of energy compared with fossil fuels. We call this process of transitioning from fossil fuels to low-carbon energy sources ‘decarbonisation’.

In the first section of this entry, we saw that our progress in decarbonising our total energy system (including transport, heat and electricity) has been slow. Fossil fuels are still the dominant energy source. If we focus on our electricity sector in particular, are we performing any better?3

Our progress over the last decade tells an interesting story which we have covered in its own blog post. These trends can be explained in the four charts which map the share of renewable, nuclear and fossil fuel sources in global electricity production. As a brief summary: over the last decade (2005-2015) the share of renewables in our electricity mix has increased by approximately 5-6 percent. This is good news. However, over this same period, the share from nuclear production has decreased by almost exactly the same amount (5-6 percent).

Overall, this means that our total share of low-carbon electricity production is almost exactly the same as a decade ago (as shown in the chart). In fact, if we compare the share of electricity produced by low-carbon sources (renewables and nuclear) in 2015 to that of 1990 , we see that it has dropped by around three percent. Progress on electricity decarbonisation has been stalled over the last decade as a result of a growing aversion to nuclear energy.

The final chart provides a breakdown of fossil fuel sources in our electricity mix. Since 2005, natural gas and coal have increased their share by one and two percent, respectively whereas the contribution from oil has declined by two percent. Nonetheless, overall, the relative mix of electricity sources has changed very little over the last few decades.

Global energy trade

The distribution of energy resources can have an obvious impact on energy trade across the world. The other important factor in energy trade is domestic levels of energy consumption. If you are a country rich in resources but also have high domestic levels of consumption, you may have little energy left to export. Similarly, if a country has low levels of energy consumption, if may still be a net exporter of energy despite have comparatively low levels of natural resources. Other influences on energy trade may be geopolitical: for example, some countries may want to converse fuel resources to maintain levels of energy security into the future.

In the two charts we have graphed energy imports and exports, both by income level and by region. Note that you can also manually select countries to compare. Here we have measured energy imports and exports as a percentage of domestic energy use, where a positive percentage indicates a country or region is a net importer of energy, and negative is a net exporter. For example, collectively high-income nations in 2014 imported nearly five percent of consumed energy.

In terms of income level, we see that there is a distinct flow of energy resources from low, middle and upper middle income to high-income nations (with the exception of lower middle income). On a continental basis, we see the dominance of energy exports from the Middle East & North Africa (being a net exporter of 127 percent of its consumption levels). Interestingly, Sub-Saharan Africa is also a net exporter of energy (despite having low levels of coal reserves and only moderate levels of oil and gas)- this is most likely a result of low levels of domestic consumption. North America and Europe & Central Asia reach approximately energy parity (effectively balancing consumption with trade). South Asia is a net importer of energy, importing approximately one-third of its energy consumption.

Investment in renewable technologies

Shifting our energy systems away from fossil fuels towards renewable technologies will require significant financial investment. But how much are we really investing in the sector, and how is this finance distributed across the world?

In the graph we see global investments in renewable technologies from 2004 to 2015 (measured in billion USD per year). In 2004, the world invested 47 billion USD. By 2015, this had increased to 286 billion USD, an increase of more than 600 percent. Investment has grown across all regions, but at significantly different rates. Note that you can use the ‘absolute/relative’ toggle on the chart to compare regions on relative terms. Growth has been greatest in China, increasing from 3 billion USD in 2004 to 103 billion USD by 2015 (an increase of 3400%). China is now the largest single investor in renewable technologies, investing approximately the same as the United States, Europe and India combined.

Combining Chinese and Indian investment with its neighbours, Asia & Oceania is the largest continental investor. Europe’s investment has been through a significant growth-peak-reduction trend, peaking in 2011 at 123 billion USD before declining to 49 billion USD in 2015. Investment in the Middle East & Africa remains relatively small, but has shown significant growth over the last ten years (after investing only 0.5 billion USD in 2004).

Levels of absolute investment tell an important story, but are disadvantaged by the fact that they take no account of the size of investments relative to a country’s economy. We might expect that the largest economies would also be the largest investors. If we want to assess which countries are making a fair ‘contribution’ or ‘share’ to investment in clean energy, it is useful to assess investment contributions as a percentage of a country’s gross domestic product (GDP). We have calculated this (as a percentage of GDP) and plotted it for the largest single-country investors in one of the charts.

This tells a slightly different story. Most countries invest less than one percent of GDP in renewable technologies (with the exception of South Africa and Chile, which make an impressive contribution at 1.4 percent). When normalised to GDP, China remains one of the largest investors, at 0.9 percent. Interestingly, despite being the second largest investor in absolute terms, the United States invested only 0.1 percent of its GDP in 2015.

Indeed, when it comes to relative contributors to renewable energy, low-to-middle income transitioning economies typically invest more than high-income nations. This may be partly explained by the fact that these nations are likely to be investing a higher percentage of their GDP into energy provision and expansion overall (whereas high-income nations typically have well-established energy systems). Nonetheless, most high-income nations have set ambitious greenhouse gas reduction targets in their commitments to the Paris climate agreement.4

Achieving these targets will require significant investments in low-carbon technologies.

We have looked at investment trends by region, but which renewable technologies are receiving the largest investment? In the chart we have shown global investment trends by energy source, through to 2016. Note that large hydropower is not included in these figures. Again, you can switch between the ‘absolute/relative’ toggle to see comparisons in each.

In 2016, solar and wind energy both received 47 percent of investment (combining to account for 94 percent of global finance). These two technologies have been taking an increasing share, especially over the last five years. In 2006, bioenergy (both in the form of biomass and liquid biofuels) took a sizable share of global investment, peaking at 36 percent. This has dwindled over the last decade, receiving less than four percent in 2016. These trends suggest that investors see solar and wind energy as the dominant renewable technologies of the future.

Energy has a crucial role to play in a global development context. The potential for energy to improve living standards, whether through the freeing of time from household chores (for example, washing clothes or cooking); increased productivity; improved healthcare and education services; or digital connections to local, regional and global networks.

The link between energy consumption and economic growth has been a topic of wide discussion. A large number of studies have attempted to derive the causal relationship between energy consumption and economic growth, however no clear consensus has emerged.5

This can be partly attributed to the fact that the link between energy and prosperity is not always unidirectional. Gaining access to electricity and other energy sources may provide an initial increase in GDP, but having higher GDP may in turn drive higher energy consumption. Additionally, progress in development outcomes can be complex: a number of parameters may be improving at the same time. If, for example, energy access and consumption, nutrition, education, health, and sanitation are all improving simultaneously (and having complex relationships with one another), it can be hard to directly attribute improvement in living standards back to a single parameter.

Chontanawat et al. (2008) carried out a systematic study across 100 countries to try to reach a common consensus on the energy-GDP link.6

Akinlo (2008) did similarly across 11 Sub-Sahara African countries to define a common relationship.7

Neither found a causal relationship which was true in all contexts. For some countries, the relationship was unidirectonal (energy consumption was a direct and long-term driver of economic growth), others are bidirectional; some are cointegrated with other factors; and for some there was actually no clear link between the two. Nonetheless, for most countries, there is an important relationship between energy and prosperity. However, the exact dynamics of each is complex and context-dependent.

What does our data suggest of this link between the two? In the chart we have plotted per capita energy consumption (on the y-axis) versus per capita GDP (PPP-adjusted) (on the x-axis) for the year 2015. Indeed, we see a strong trend: typically the higher a country’s average income, the more energy it consumes. In our second chart we present the percentage of the population with access to electricity (y-axis) versus GDP per capita (x-axis). If we press play and watch how these trends evolve through time, we see a similar trend: both electricity access and prosperity increase for most countries through time. However, in both of these visualisations, it’s challenging to differentiate how much of this trend can be explained by energy-led growth, and how much is a result of growth-led energy consumption.

In our final chart we have plotted the relationship between per capita energy consumption (y-axis) and the share of the population in extreme poverty (x-axis). In general, we see a trend of poverty alleviation with higher energy consumption levels. However, this does not necessarily hold as a direct relationship for all countries.

Relative cost of energy sources

Prices can strongly influence our choice of energy sources. In this regard, it is the relative cost between sources which is important. This is true in higher-income countries (we want low energy bills), but is increasingly important in low and middle-income economies. For many countries, increasing the share of the population with access to electricity and energy resources is a key priority, and to do so, low-cost energy is essential.

How do we compare the relative cost of energy? The dominant energy source in the transport sector is liquid fuels (diesel and gasoline) for which relative costs are less important than changes in price through time. Let’s therefore focus on the relative costs of energy sources in the electricity sector.

To do this, we compare costs based on what we call the ‘levelised cost of electricity’ (LCOE). The LCOE attempts to provide a consistent comparison of electricity costs across sources but taking the full life-cycle costs into account. It is calculated by dividing the average total cost to build then operate (i.e. both capital and operating costs) an energy asset (for example a coal-fired power station, a wind farm, or solar panel) by the total energy output of that asset over its lifetime. This gives us a measure of the average total cost per unit of electricity produced. Measuring sources on this consistent basis attempts to account for the fact that resources vary in terms of their capital and operating costs (for example, solar PV may have higher capital costs, but lower operating costs relative to coal over time). Note that this cost of energy production has an obvious impact on electricity prices for the consumer: the LCOE represents the minimum cost producers would have to charge consumers in order to break-even over the lifetime of the energy project.

To be truly competitive, renewable technologies will have to be cost-competitive with fossil fuel sources. In the chart, sourced from IRENA’s latest Rethinking Energy report, we see the LCOE (measured in 2016 USD per megawatt-hour of electricity produced) across the range of renewable technologies in 2010, and in 2016.8

It’s important to acknowledge that the relative costs of energy are context-dependent and vary across the world. For example, the relative cost of solar PV is likely to be lower in lower latitude countries than at high-latitudes because they will produce more energy of their lifetime. This can produce very different LCOE figures by region (and indeed the country-specific LCOE charts can vary significantly). For our global chart, this range of costs is represented as vertical bars for each technology. The white line in each represents the global weighted average cost per technology.

Similarly, the cost of fossil fuels can very depending on the fuel quality, ease of extraction and regional resources. The average range of fossil fuel costs is shown as the grey horizontal block.

What we see is that in terms of the 2016 weighted average cost, most renewable technologies are within a competitive range of fossil fuels. The key exception to this is solar thermal which remains about twice as expensive (although is falling). Hydropower, with the exception of traditional biomass, is our oldest and well-established renewable source: this is reflected in its low price (which can undercut even the cheapest fossil fuel sources). Note however that although the weighted average of most sources is competitive with the average fossil fuel cost, the wide range of potential costs means that this is not true for all countries. This is why the selection of particular technologies need to be considered on a local, context-specific basis.

If we consider how the average cost of technologies changed from 2010-16, we see that both solar PV (and to a lesser extent, solar thermal) dropped substantially. This cost reduction in solar PV has been dramatic over the past few decades, as shown in the chart. The price of solar PV modules has fallen more than 100-fold since 1976. On average, the technology has had a learning rate of 22 percent; this means that the cost falls by 22 percent for every doubling in solar PV capacity (although progress has not necessarily been constant over this period).

Levelised cost of electricity (LCOE) 2010 and 20169
Lcoe 2010 2016
Solar pv prices vs cumulative capacity v11 850x600
Click to open interactive version

What are the safest sources of energy?

Two centuries ago we discovered how to use the energy from fossil fuels to make our work more productive. It was the innovation that started the Industrial Revolution. Since then, the increasing availability of cheap energy has been integral to the progress we’ve seen over the past few centuries. It has allowed work to become more productive, and people in industrialized countries are much richer than their ancestors, work much less, and enjoy much better living conditions than ever before. Energy access is therefore one of the fundamental driving forces of development. The United Nations rightly says that “energy is central to nearly every major challenge and opportunity the world faces today.” 

But while energy from fossil fuels brought many benefits it unfortunately also has major negative consequences. There are three main categories of negative consequences.

The first is air pollution: at least five million people die prematurely every year as a result of air pollution.10.  

Fossil fuels and the burning of biomass – wood, dung, and charcoal – are responsible for most of those deaths. Eliminating fossil fuels could cut premature deaths from air pollution by around two-thirds. That’s three to four million deaths per year.11

The second is accidents. This includes accidents that happen in the mining and extraction of the fuels (coal, uranium, rare metals, oil and gas) And it includes accidents that occur in the transport of raw materials and infrastructure, the construction of the power plant; or their deployment. 

The third is greenhouse gas emissions: fossil fuels are the main source of greenhouse gases, the primary driver of climate change. In 2018, 87% of global CO2 emissions came from fossil fuels and industry.12

All energy sources have negative effects. But they differ enormously in size: as we will see, in all three aspects, fossil fuels are the dirtiest and most dangerous, while nuclear and modern renewable energy sources are vastly safer and cleaner.

From the perspective of both human health and climate change, it matters less whether we transition to nuclear power or renewable energy, and more that we stop relying on fossil fuels.

Nuclear energy and renewables are far, far safer than fossil fuels

Today the global energy system is still dominated by fossil fuels, traditional biomass, hydropower and nuclear energy.13 However, modern renewables, such as solar and wind, are growing and we expect them to play an increasing role in our energy systems in the coming decades. As our energy systems transition we have decisions to make about what sources to choose. Safety concerns should be a key factor we consider.

How do fossil fuels, nuclear energy and renewables stack up in terms of safety?

Research that answers this question comes from several sources. Anil Markandya and Paul Wilkinson (2007) published an analysis in the medical journal The Lancet, which compared the death rates from fossil fuels, nuclear, hydropower and biomass.14 In this study they considered deaths from accidents – such as the Chernobyl nuclear disaster, occupational accidents in mining or power plant operations – as well as premature deaths from air pollution.15 When they published the paper, modern renewable energy sources were still a very small source of energy production and weren’t included in the analysis. Data on the safety of renewable sources was published in a study by Benjamin Sovacool and colleagues (2016).16,17 We’ve combined the results of these studies so we can compare death rates from all energy sources.

Later in this article we discuss in more detail how death rates from nuclear energy were derived. But to summarise: the death rate for nuclear includes an estimated 4000 deaths from the 1986 Chernobyl disaster in Ukraine (based on estimates from the WHO); 574 deaths from Fukushima (one worker death, and 573 indirect deaths from the stress of evacuation); and estimated occupational deaths (largely from mining and milling), as provided by Markandya and Wilkinson (2007).

In the chart we see the death rates of each – given as the number of deaths per terawatt-hour of energy.18 One terawatt-hour of energy is about the same as the annual energy consumption of 27,000 citizens in the European Union.19

We see massive differences in the death rates of nuclear and modern renewables compared to fossil fuels.

Nuclear energy, for example, results in 99.8% fewer deaths than brown coal; 99.7% fewer than coal; 99.6% fewer than oil; and 97.5% fewer than gas. Wind, solar and hydropower are more safe yet.

Putting death rates from different energy sources in perspective

Looking at deaths per terawatt-hour can seem a bit abstract. So let’s try to put it in perspective.

Let’s consider how many deaths each energy source would cause for an average town of 27,000 people in Europe, which – as I’ve said before – consume one terawatt-hour per year. Let’s call this town ‘Euroville’.

If Euroville was completely powered by coal we’d expect 25 people to die prematurely every year as a result. Most of these people would die from air pollution). This is how a coal-powered Euroville would compare with towns powered by other energy sources:

  • Coal: 25 people would die prematurely every year;
  • Oil: 18 people would die prematurely every year;
  • Gas: 3 people would die prematurely every year;
  • Nuclear: In an average year nobody would die. A death rate of 0.07 deaths per terawatt-hour means it would take 14 years before a single person would die. As we will explore later, this might even be an overestimate.
  • Wind: In an average year nobody would die – it will take 29 years before someone died;
  • Hydropower: In an average year nobody would die – it will take 42 years before someone died;
  • Solar: In an average year nobody would die – only every 53 years before someone would died.

Fossil fuels kill many more people than nuclear energy. We’ve seen this from the comparison above. This is surprising to many people, because many have prominent memories of the two major nuclear disasters in history: Chernobyl and Fukushima.

Unfortunately, public opinion on nuclear energy tends to be very negative.

In a separate article we take a look at the death toll of Chernobyl and Fukushima in detail. But we should at this stage summarise how the death rates for nuclear energy were calculated.

When we try to combine the two analyses referenced earlier, one issue we encounter is that neither study includes both of the major nuclear accidents in its death rate figure: Markandya and Wilkinson (2007) was published before the Fukushima disaster in 2011; and Sovacool et al. (2016) only look at death rates since 1990, and therefore do not include the 1986 Chernobyl accident. We have therefore reconstructed the death rate for nuclear to include both of these terrible accidents.20

For Chernobyl, there are several death estimates. We rely on the estimate published by the World Health Organization (WHO) – the most-widely cited figure – although this is considered to be too high by several researchers, including a later report by the United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR).21 We discuss this contention below. The WHO estimates that 4000 people have, or will die, from the Chernobyl disaster. This includes the death of 31 people as a direct result of the disaster and those expected to die at a later date from cancers due to radiation exposure.

The disaster in Fukushima killed 574 people. In 2018, the Japanese government reported that one worker has since died from lung cancer as a result of exposure from the event. No one died directly from the Fukushima disaster. Instead, most people died as a result of evacuation procedures. According to Japanese authorities 573 people died due to the impact of the evacuation and stress.22

To the death toll of history’s two nuclear disasters we have added the death rate that Markandya and Wilkinson (2007) estimated for occupational deaths, most from milling and mining. Their published rate is 0.022 deaths per TWh.

The sum of these three data points gives us a death rate of 0.07 deaths per TWh. We might consider this an upper estimate. Our estimated death toll from Chernobyl is based on the 2005/06 assessment from the WHO which applies a very conservative methodology called the linear no-threshold model. If you’re interested in the details of this we discuss it in more detail here. A later report by the United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR) suggests that this overstates the risk of radiation-related deaths.23

Even with this upper figure, nuclear is still much much lower than the death rate from fossil fuels – 350 times lower than coal. Despite this, politicians have turned their backs on it in many countries.

Modern renewables and nuclear energy are not only safer but also cleaner than fossil fuels

So far we’ve only considered the short-term health impacts of these energy sources. But we need to also take into account their longer-term impact on climate change.

There is actually very good news: the safest sources for us today are the same sources that have the smallest impact on the climate. Sometimes the solutions to the large global issues we face come with trade-offs, but not here. Whether you are concerned about people dying now or the future of the planet, you want the same sources of energy.

The visualization here shows this. On the left I have plotted the death rates per unit energy data we looked at previously and on the right you see their greenhouse gas emissions per energy unit .This measure of greenhouse gas emissions considers the total carbon footprint over the full lifecycle; figures for renewable technologies, for example, take into consideration the footprint of the raw materials, transport and their construction. I have adopted these figures as reported in the IPCC’s 5th Assessment Report, and more recent life-cycle figures by Pehl et al. (2017), published in Nature.24,25,26

The world is not facing a trade-off – the safer energy sources are also the least polluting. We see this from the symmetry of the chart. Coal causes most harm on both metrics: it has severe health costs in the form of air pollution and accidents, and emits large quantities of greenhouse gas emissions. Oil, then gas, are better than coal, but are still much worse than nuclear and renewables on both counts.

Nuclear, wind, hydropower and solar energy fall to the bottom of the chart on both metrics. They are all much safer in terms of accidents and air pollution and they are low-carbon options.

Unfortunately they still account for a very small share of global energy consumption – less than 10% of primary energy. Each source’s share of global primary energy production in 2019 (including traditional biomass in the total) is shown in the centre. Fossil fuels have so far dominated our energy systems for a couple of reasons: they kickstarted the Industrial Revolution and since then much of our energy infrastructure has been built around them. This early investment in fossil fuels means they have for a long time been relatively cheap – cheaper than many modern renewables in their infancy. But today, if we factor in the total costs of fossil fuels – not only the energy costs but also the social cost to our health and the environment – they are much more expensive than the alternatives. If we were to impose a carbon tax – which would account for the total costs that we all suffer – this would be the case.

Fortunately, clean and safe renewable technologies are becoming economically-competitive in their own right. The market price of both solar and wind has been falling rapidly meaning there is a real chance for change.

There is fierce debate about which low-carbon energy technologies we should pursue. But on the basis of three key questions – human health, safety and carbon footprint – nuclear and modern renewables do clearly best. A number of studies have found the same: there are large co-benefits for human health and safety in transitioning away from fossil fuels, regardless of whether you replace them with nuclear or renewables.27

The air pollution that fossil fuels cause is killing millions of people every year, and endanger many more from the future risks of climate change. We must shift away from them. And we can, we have better alternatives.

What is the safest form of energy

Data Sources

In this section

Long run

The History Database of the Global Environment (HYDE)

Correlates of War

  • Data: ‘Primary Energy Consumption’ (thousands of Metric-ton Coal Equivalent)
  • Geographical coverage: Global – by country
  • Time span: 1816-2007
  • Available at: Online at
  •  The original data source for data after 1970 is the United Nations’ Energy Statistics Database UNESD. Data for the time before 1970 is taken from Mitchell International Historical Statistics. See the codebook on page 61.

The Shift Project (TSP)

  • Data: Historical Energy Consumption Statistics and Historical Energy Production Statistics
  • Geographical coverage: Global – by country and world region
  • Time span: Since 1900
  • Available at: Both datasets are online at

Post 1950

IEA – International Energy Agency

  • Data: Data on electricity, oil, gas, coal and renewables. Data on CO2 emissions (also projections)
  • Geographical coverage: Global – by country
  • Time span: Last decades
  • Available at: Online at
  • The IEA is publishing the World Energy Outlook.
  •  You have to pay to access the IEA databases. But some data is available through Gapminder, for example Residential Energy Use (%). (for few countries since 1960, for more countries since 1971 or 1981)

Energy Information Administration

  • Data: Total and crude oil production, oil consumption, natural gas production and consumption, coal production and consumption, electricity generation and consumption, primary energy, energy intensity, CO2 emissions and imports and exports for all fuels
  • Geographical coverage: Global – by country
  • Time span: Annual data since 1980
  • Available at: Online at
  •  EIA is a US government agency.

BP Statistical Review of World Energy

  • Data: BP publishes data on Oil, Gas Coal, Nuclear Energy, Hydroelectricity, Renewables, Primary Energy Consumption, Electricity Generation, Carbon Dioxide Emissions
  • Geographical coverage: Global – by country and region
  • Time span: Annual data since 1951
  • Available at: Online at

World Development Indicators – World Bank