The energy transition will be much cheaper than you think
Most analysts overestimate energy demand and underestimate technological advances
Video: Jiaqi Wang
People who want to do more to fight climate change and those who want to do less tend to have one thing in common. Both sides agree that decarbonising the world economy will be dauntingly expensive. At this week’s annual UN climate summit, in Baku, Azerbaijan, the numbers being bandied around are in the tens of trillions of dollars.
Many see such spending as a colossal waste. Donald Trump, America’s president-elect, denounced the Paris agreement to cut global emissions, reached at the 2015 climate summit, as something that “hurts Americans, and cost a fortune”. He withdrew America from it in his first presidency. Because America has since rejoined, he is likely to do so again. Climate activists, for the most part, do not dispute the hair-raising price tag; they simply consider the expense worthwhile when weighed against the catastrophic damage unchecked climate change is likely to inflict.
Yet this one point of agreement between climate activists and carbon addicts is, in fact, wrong. Greening the world economy will be much cheaper than the two groups imagine. The Economist has looked at estimates of the global cost of an “energy transition” to a zero-emissions world from a range of economists, consultants and other researchers—the sort of estimates that routinely form the basis for policymaking. They range from around $3trn a year to almost $12trn a year, which is indeed a lot. But these figures are overblown in four important ways.
Bakuky assumptions
First, the scenarios being costed tend to involve absurdly speedy (and therefore expensive) emissions cuts. Second, they assume that the population and economy of the world, and especially of developing countries, will grow implausibly rapidly, spurring pell-mell energy consumption. Third, such models also have a record of severely underestimating how quickly the cost of crucial low-carbon technologies such as solar power will fall. Fourth and finally, the estimates disgorged by such modelling tend not to account for the fact that, no matter what, the world will need to invest heavily to expand energy production, be it clean or sooty. Thus the capital expenditure needed to meet the main goal set by the Paris agreement—to keep global warming “well below” 2°C—should not be considered in isolation, but compared with alternative scenarios in which rising demand for energy is met by dirtier fuels.
The incremental bill to cut emissions is likely to be less than $1trn a year, which is to say less than one percent of global GDP—not peanuts, but not an unaffordable pipe dream, either. That may sound optimistic, but it is probably still an overestimate, since it corrects only for the fourth flaw in most estimates: the failure to account for the cost of business as usual. Slower economic growth, cheaper technology and more modest targets for when the world reaches net zero could reduce the price tag even further.
According to the International Energy Agency (IEA), a think-tank for rich countries, roughly $3trn, or 3% of global GDP, has been invested in energy in 2024. That is a record, fuelled partly by cyclical investment in oil and gas and partly by rising investment in clean power generation, which was level in the 2010s, but has been growing since. Roughly three-quarters came from private sources and a quarter from governments, in line with the recent trend.
The recipients of that investment, though, have changed profoundly since the Paris agreement. In 2015 less was being invested in clean technology than in fossil fuels. Today clean technology receives almost twice as much. This year solar power should account for $500bn, more than every other source of generation combined.
These numbers flatter clean energy somewhat, since they include investments in electric vehicles (EVs), heat pumps and improvements to electric grids, which do not in themselves lower emissions that much. Instead, they pave the way for big emissions cuts, provided that the electricity used comes from low-carbon sources. The spread of EVs in China, for instance, reduces global oil demand but makes only a small contribution to reducing emissions since the vehicles’ batteries are charged from China’s coal-heavy grid.
Nonetheless, the outlook for the climate is improving. In 2015 the “Emissions Gap Report” the UN Environment Programme (UNEP) produces before every climate summit projected that, on the basis of policies then in force around the world, global average temperatures would be almost 5°C higher than in pre-industrial times by the end of the century. This year’s report puts that number at just over 3°C. Other forecasters are even more optimistic: the IEA reckons current policies will yield around 2.4°C of warming. Bloomberg New Energy Finance (BNEF), a research outfit, thinks existing policies and the falling prices of green technologies will lead to 2.6°C of warming by 2050. Wood Mackenzie, a consultancy, is forecasting 2.5°C by 2100 as its base case.
None of these projections, however, imagine that the world will keep warming below 2°C, as the Paris agreement stipulates, let alone below 1.5°C, the supplementary target that signatories said they would try to meet. There is a wide range of views about how much investment is needed to meet these goals. Naturally, though, staying below 1.5°C is costlier than staying below 2°C. It is the cost of the 1.5°C target that typically gets the most attention.
To estimate costs, economists combine a model of the economy and a scenario that represents the achievement of a given goal. This could be a temperature target, such as the “pathways” to 1.5°C or 2°C laid out by the Intergovernmental Panel on Climate Change (IPCC), a UN body. Or it could be a target for the global volume of emissions at a given time. The IEA’s net-zero scenario assumes that, by the middle of the century, all greenhouse gases pumped into the atmosphere will be offset by equivalent removals. There is a tendency to see net zero by 2050 as being roughly equivalent to meeting the 1.5°C goal, although modellers normally allow a brief overshoot in temperature, which falls back as the removal of carbon from the atmosphere gathers pace.
The IEA’s modelling finds that reaching net zero by 2050 will require $5trn a year of investment in clean energy by 2030. That is more than twice the $2trn a year it reckons is currently going into clean energy and two-thirds more than its estimate of total current investment in energy. A similar scenario from BNEF involves $5.4trn a year this decade. McKinsey Global Institute, a research outfit, puts the annual cost of net zero by 2050 at $9.2trn; Wood Mackenzie at just under $3trn. UNEP estimates that a range of $7trn to $12trn per year will be needed by 2035 to limit warming to 1.5°C.
What money cannot Dubai
This wide divergence stems from different modelling methodologies. Whatever your approach, though, applying any model to near-impossible scenarios yields suspect results. And limiting warming to 1.5°C is, broadly speaking, impossible. The Global Carbon Budget, a consortium of scientists, estimates that temperatures will permanently reach that level in six years at the current rate of emissions. Preventing any further climate change would entail ending all greenhouse-gas emissions within that time—a prohibitively expensive, if not impossible, task.
Keeping global warming below 2°C is much more plausible, happily. The Global Carbon Budget estimates that it will take 27 years for the world to hit that increase in temperatures at the current rate of emissions. The big expansion of breathing room, in turn, allows for a slower and therefore cheaper transition.
Yet much analysis remains focused on more stringent goals. That is natural. The inclusion of an aspiration of 1.5°C in the Paris agreement was considered a great victory by the most vulnerable countries and by climate activists. Three years later the IPCC produced a vast report showing that even 1.5°C of warming would be very damaging, and that 2°C would be catastrophic for many countries and ecosystems. The extent and severity of the harm rises inexorably with the temperature. But when trying to decide what to do, it is of little help to demonstrate that achieving the impossible is impossibly expensive.
Another problem with the models is their assumptions about economic growth. Matt Burgess of the University of Wyoming and colleagues note that the IPCC’s projections have tended to overestimate economic growth in both the rich and poor worlds. They suggest that the worst case for economic growth among the “Shared Socioeconomic Pathways” (SSPs) the IPCC uses in its modelling is in all likelihood more of a best-case scenario. They forecast GDP per person based on the historical relationship between its absolute level and its rate of growth. That yields much lower projections than SSP2, supposedly a “middle of the road” scenario (see chart 1).
Even the IEA’s assumption of 2.7% annual average global growth until 2050, although in line with recent experience, may in the end prove optimistic. It is based on projections of population growth from the UN that have consistently failed to foresee drops in birth rates in the developing world. Fewer people means lower economic growth, all other things being equal. And an older planet, with even fewer people of working age, is likely to grow more slowly.
Just as relaxing the temperature target leads to big cost reductions, so does lowering the demand for energy as a result of slower economic growth. And just as with missing the 1.5°C target, this is not really a good thing. A world with lower growth is a bad one in many ways, for the poor in particular. If higher growth could somehow be arranged, especially in the poorest countries, that would be a boon for the world, even if it meant that more money would have to be spent on decarbonisation. But basing estimates of the cost of decarbonisation on wishful thinking about growth rates makes them unduly expensive. To get an accurate picture, better to be realistic.
Economic modellers also have a poor record of predicting technological advances. They overestimate the take-up of some technologies (such as carbon capture and storage, whereby carbon dioxide is sucked out of the smokestacks of power stations and factories and stashed away safely underground) and severely underestimate the falling costs of others, most notably solar panels and lithium batteries. Rupert Way of the University of Cambridge and others have modelled an energy system in which the cost of solar power, wind power, lithium batteries and hydrogen electrolysers falls according to “Wright’s law”. This holds that every doubling of production sees unit costs fall by a fixed percentage, with that percentage derived from past experience. In this scenario emissions fall so rapidly that even the 1.5°C target can be met at minimal cost.
In practice bottlenecks always form in fast-growing industries, impeding the spread of new technologies despite falling costs. As cheap as solar power has become, for instance, securing grid connections for it remains a slow process in many countries. By the same token, there are fewer than two dozen ships outside China capable of installing an offshore wind farm. All of them, unsurprisingly, are booked up for years in advance. Modellers try to reflect these obstacles by placing arbitrary limits on how quickly the cost of new technologies can fall. But they have tended to apply these brakes too heavily, especially for renewable power. The IEA’s predictions of renewable generation capacity have repeatedly fallen wildly short over the past decade (see chart 2).
Another factor that exaggerates the cost of decarbonisation is the failure to consider the counterfactual in which decarbonisation does not take place. Wood Mackenzie has devised a “delayed transition” scenario, in which trade tensions and geopolitical strife lead countries to hold back on their transition to a zero-carbon energy system. This leads to 3°C of warming. But it still entails $52trn of investment in the energy system by 2050. The same consultancy’s estimate for the cost of getting to 2°C is $65trn.
In other words, the cost in terms of energy investment of doing almost nothing about global warming is not that much lower than the cost of limiting global warming to 2°C. The additional $13trn that Wood Mackenzie thinks would be required over 25 years translates to roughly 0.5% of current global GDP a year—and less as the world economy grows. This is in broad agreement with a paper that David McCollum, a climate scientist, and others published in 2018. That put the incremental cost of decarbonising the energy system to meet a 2°C goal at $320bn a year, which is equivalent to $400bn today. Even UNEP’s estimate of a $7trn-12trn annual cost to meet a 1.5°C target is reduced to between $900bn and $2.1trn once investment that would happen anyway is excluded. It would fall even further if less expansive assumptions were used about future economic growth.
There is a catch: the timing of the necessary investments is not the same in a low-carbon world as in a grimy one. Business-as-usual scenarios tend to assume that investment will be spread out roughly evenly across the period under consideration. The strictures on cumulative emissions that a 2°C carbon budget involves mean that more investment in clean energy is needed earlier in the forecast period. The Energy Transitions Commission, an industry initiative, reckons that total clean-energy investment must quadruple from around $1trn in 2020 to $4trn in 2040 before falling back again. Investment in fossil fuels will decline on a similar trajectory, reducing the net cost and, eventually, leading to operational savings from the much lower demand for fossil fuels.
Sharm reduction
But even assuming the costs are front-loaded, the tab for reaching 2°C need not be overwhelming. And though 1.5°C is not achievable, the models also suggest that spending more now might put Earth on a path to 1.8°C of warming or less. Reducing overall warming by a few tenths of a degree might in fact pay for itself, insofar as the world would suffer less damage in total from global warming.
Three problems could yet blight this rosy outlook. The first is that, although decarbonising power generation and transport is the most important element in mitigating climate change, it is not the only one. There is also agriculture, which is a big source of greenhouse gases other than carbon dioxide, such as methane and nitrous oxide. The technologies that might help reduce these emissions are much less established. It is therefore much harder to make any confident predictions about the future cost of curbing these emissions.
Mismatched incentives are another problem. The people who will suffer the most harm from global warming are not the people in the best position to pay to curb it. Poorer countries need more investment but cannot afford it.
This is made worse by the cost of capital. Most climate scenarios have historically assumed a single cost of capital for the whole global economy. But the poorer countries where most is at risk face a higher cost of capital than richer ones. The Climate Policy Initiative, a think-tank, calculates that investors in a solar farm in Germany need a return of 7% on the capital invested to break even, given typical borrowing costs. In Zambia prohibitive lending rates for business raise the necessary return to 38%. Unless financing costs in the developing world can be reduced, the price tag for decarbonisation will rise.
The final caveat is that models, almost by their nature, tend towards the rational. Policy is less reliable in this respect. Things which should be affordable are often in practice exorbitant because of incompetent delivery, constraints imposed by other political goals and graft.
Most models assume that society will try to complete the energy transition as cheaply as possible. Yet that definitely will not happen. Many governments feel the need to rule out some useful techniques to lower the cost, such as carbon taxes, and adopt unnecessarily expensive methods instead, such as subsidising the manufacture of emissions-cutting technology to help boost their industrial base. There is often a political imperative to assuage mining lobbies or fossil-fuel-rich regions, or to protect manufacturers unable to compete with cheaper foreign producers of batteries, electric vehicles or solar panels.
Sometimes there is a tussle over how to spend whatever money politicians set aside for climate, in which preparing for climate change competes with curbing it. This is a genuinely hard trade-off, akin to a prisoners’ dilemma. The less the world as a whole spends on decarbonisation, the more rational it is in any given country to spend a bigger proportion of the climate budget on adaptation, not mitigation.
As important as these notes of caution are, however, they do not alter the fact that the cost of a transition away from fossil fuels is consistently exaggerated. This is no coincidence: climate sceptics and climate activists both have reason to talk up the expense. The sceptics can use alarming numbers as a reason not to bother; the activists can deploy them to demand more spending. In fact, climate change is neither the end of the world nor an expensive hoax. It is a real and difficult problem, but one that can be curbed affordably. ■
For more coverage of climate change, sign up for the Climate Issue, our fortnightly subscriber-only newsletter, or visit our climate-change hub.
This article appeared in the Briefing section of the print edition under the headline “Carbon bargain"
