Freeing Sisyphus: new rules of thumb for policymaking on decarbonisation

By Simon Sharpe, Director of Economics, UNFCCC Climate Champions; Policy Associate, Institute for Innovation and Public Purpose, University College London

Every day, governments around the world make laws, regulations, taxes, and investments that influence future emissions of greenhouse gases.  While public debate on climate change often focuses on countries’ national emissions targets, it is these specific policies that determine whether targets are missed, met, or exceeded. 

Progress so far has been underwhelming, to put it mildly.  Since the international agreement thirty years ago to ‘prevent dangerous anthropogenic interference with the climate system’, annual global emissions of greenhouse gases have continued to rise.  To have a reasonable chance of meeting the goals of the Paris Agreement – limiting global warming to below 1.5 degrees C – we now need to decarbonise the global economy roughly five times faster over the course of this decade than we managed during the last two decades.  Experts are beginning to say that this is implausible.  Meanwhile, scientists are telling us that the risks of catastrophic changes are greater than they thought, even at low levels of warming.

In this context, it matters greatly how the policy decisions that influence emissions are made.  Most are considered to be questions of economic policy, since they affect economic interests and involve the reallocation of resources.  They are typically informed by economic analysis – whether formally, through modelling and calculations, or informally, through the application of rules of thumb based on widely-accepted theory.  

In a recent study of outstanding examples of success in decarbonisation in China, India, Brazil and Europe, researchers found that the policies most critical to these successes were generally implemented ‘despite, not because of, the predominant economic analysis and advice.’  This must surely be cause for concern.  I argue here that a different approach to economic decision-making is possible, one that will help governments act more effectively, more often.  This has implications not only for national policy, but also for international cooperation.  

Changing the foundational assumption 

Much of the economic analysis and advice that informs climate change decision-making is based, whether consciously or not, on the assumption of equilibrium.  In economics, equilibrium is defined as ‘a situation in which nobody has any immediate reason to change their actions, so that the status quo can continue, at least temporarily.’  This foundational assumption underlies the structure of economic models, the design of decision-making frameworks such as cost benefit analysis, and the formation of normative rules of thumb for policymaking. 

This assumption contrasts markedly with the challenge we face.  Meeting climate change goals requires rapid and deep ‘system transitions’ in the global economy in each of the emitting sectors.  A system transition is anything but a continuance of the status quo.  It involves the creation and spread of new technologies, markets, business strategies, infrastructure, institutions, and cultural norms – providing many reasons for people to change their actions.  If system change is our goal, then by definition we are dealing with the economy in disequilibrium.  That must be our new starting assumption.

A body of theory that can describe, explain and predict the behaviour of the economy in disequilibrium already exists, though it is not yet widely or systematically applied to policy.  Disequilibrium models have been built, although there are few instances yet of their being put to use by governments.  Empirical evidence of how decarbonisation policies work is accumulating, in many cases contradicting traditional assumptions.  From this relatively new body of theory, models, and observational evidence that apply to the economy in disequilibrium, I attempt here to describe a new set of rules of thumb for policymaking on decarbonisation.  

Rules of thumb are important because they influence what questions we ask; what options we compare; what evidence we seek; and what models we build.  They help us decide, in situations of uncertainty.  They shape the views of those whose support for a decision may be needed, but who may not have time to consider the evidence in detail themselves.  Their influence on the policymaking process is profound.  We need to ensure this is a helpful influence, and not one that is actively unhelpful.  

Rule 1: Focus on feedbacks 

A common approach to decision-making in contexts of equilibrium is to compare expected outcomes of different policies at one or several fixed points of time in the future.  In contexts of disequilibrium – where our aim is to effect change in the economy – this ‘comparative statics’ approach is insufficient.  We need to understand how policies will affect processes of change in the economy: in which direction they will drive change; with what magnitude, and at what pace.  

The behaviour of a complex system such as the economy in disequilibrium can be understood by analysing its feedback loops.  Reinforcing feedbacks create self-amplifying change; an example in the climate system being how higher temperatures result in less Arctic sea ice, causing less sunlight to be reflected, leading to further heating.  Balancing feedbacks are self-limiting, tending to slow or prevent change; a thermostat in the home is the classic example.  If our aim is to drive rapid change in the economy, we should look to create reinforcing feedbacks that accelerate change in our desired direction (towards zero emissions), and to break any balancing feedbacks that stand in our way.

A simple example of where feedbacks have often been overlooked is in carbon pricing policy.  Traditionally, it has been thought that the two alternative forms of carbon pricing – a tax, and a cap-and-trade scheme – are equivalent in their economic efficiency, differing only in the detail of their implementation.  Governments have often chosen to implement cap-and-trade systems, and activists have urged them to do so, because this approach offers certainty about how much emissions will be reduced over a given period of time.  However, as the climate scientist James Hansen first pointed out, a cap-and-trade system creates a balancing feedback: if one actor finds a way to reduce emissions, this reduces demand for emissions permits; since the supply of permits is fixed, this reduces the price of a permit, and so reduces the incentive for other actors to reduce their emissions.  A carbon tax has no such self-limiting effect: if it causes me to reduce my emissions, that does not lessen the incentive for you to reduce yours.  Consistent with this understanding, a simulation of the alternative policies using a disequilibrium model shows that a carbon tax can reduce emissions more quickly, and at lower cost, than a cap-and-trade scheme.  

The next rule shows how a focus on feedbacks could have large consequences over the course of time.

Rule 2: Targeted investment is more efficient than carbon pricing  

Traditionally, carbon pricing has been thought to be the most efficient policy for decarbonisation.  This has been reflected in advice to governments from the World Bank, the International Monetary Fund, and many authoritative economists.  

In the study of successful examples of decarbonisation mentioned above, the policies that made the most important contributions were those that involved targeted investment in the deployment of new technologies.  The growth of solar power, once seen as infeasibly expensive but now hailed as providing ‘the cheapest electricity in history’, was driven largely by subsidies, notably first in Germany, and later in China.  Similarly, Brazil used subsidy and cheap public finance to grow the share of wind in its power generation mix more quickly than any of the other large emerging economies, while creating an industry supporting 150,000 jobs.  The UK used subsidised fixed-price contracts to bring down the cost of offshore wind by 70% within a decade.  India used massive public procurement to drive a transition to efficient LED lighting, bringing down its costs by over 90% in less than a decade, while increasing its deployment by a factor of several hundred and bringing electric lighting to hundreds of millions of households for the first time.

Stephane Hallegatte and Julie Rozenberg, two of the World Bank’s leading economists on climate change, wrote in 2019 that ‘Today, renewable energy is cheaper than coal in many places in the world, all major car manufacturers are working on several electric car models, and cities are starting to switch to electric buses.  All of this was achieved with policies focussed on new investments, not with carbon taxes.’  

From a disequilibrium point of view, this is not surprising.  The development and diffusion of new technologies is driven by reinforcing feedbacks – processes of self-amplifying change.  These include learning-by-doing (the more something is made, the better it can be made), economies of scale (the more it is made, the more cheaply it can be made), and the emergence of complementary technologies (the more something is used, the more technologies emerge that make it more useful).  It is these reinforcing feedbacks that can lead to exponential growth in the market share of a new technology, in the early stages of a transition.  Targeted investment in the deployment of new technologies channels economic resources towards them, directly creating or strengthening these feedbacks.  Early in a transition, a carbon price does not have the same effect: the pressure it applies can most easily be absorbed by operating fossil fuelled systems more efficiently, without directing any resources to the creation of a new system.  Consequently, in terms of its dynamic effect, investing in the new is a more efficient approach than taxing the old.  

Like any rule of thumb, this one does not always hold.  Once the new technologies have become competitive with the old, a combination of tax and subsidy can be highly effective at tipping the scales in favour of the new.  Once an old technology such as coal power has been reduced to a minority share of the market, a tax may be an effective way to activate the reinforcing feedbacks of its destruction (where divestment raises costs, leading to further divestment).  Examples of these exceptions are given under rule 4. 

Rule 3: Technology choice should be deliberate, not accidental 

A traditional principle for decarbonisation policy has been that of ‘technology neutrality’.  A ‘technology neutral’ policy is thought to allow the market to decide which technology best provides the desired function, tending to maximise economic efficiency.

In a disequilibrium economy, there is no such thing as technology neutrality.  Change in the economy is path dependent: past decisions constrain current choices, and current choices affect future options.  Every decision will, unavoidably, advantage some technologies over others, changing the shape of the future economy.  For example, a government that wishes to incentivise the purchase of zero emission vehicles may decide that as a ‘technology neutral’ policy, it will offer consumers the same subsidy when they buy a battery electric car or a hydrogen fuel cell car.  However, the effect of this policy on the two technologies will not be the same.  The battery electric technology starts at an advantage: electricity has been widely used as an energy carrier for the last century whereas hydrogen has not; consequently, electric charging infrastructure is more readily available than hydrogen refuelling infrastructure.  Giving the two technologies equal subsidy is, de facto, a decision to maintain the dominance of the one that is already ahead.  

If neutrality is impossible and choice is inevitable, it is surely preferable to choose deliberately rather than accidentally.  In our example, the two options have different pros and cons.  A transition to battery electric vehicles could support the decarbonisation of the power sector, by allowing millions of car batteries to store and release energy in a way that helps balance supply and demand in the electricity grid.  The hydrogen option could instead support the decarbonisation of industry, by bringing down the cost of hydrogen technologies through mass production.  A third option for low emission vehicles, biofuels, would be less disruptive for the car industry, but would cause higher levels of local air pollution and would compete with agriculture for the use of land.  Each of these options could give a country better or worse prospects for developing its own car industry and exporting to the global market, depending on the technology choices of other countries.  

Often the right technology choice will not be obvious, and there may be advantages in experimenting with different options, particularly in the early stages of a transition.  But even then, these options will have to be chosen.  The illusion of neutrality is a dangerous one.  An unconscious choice is most likely to be a choice in favour of incumbent technologies – those that are already ahead.  When the aim of policy is system change, that may be the opposite of what is needed.  

Rule 4: Regulation can reduce costs 

In an equilibrium economy where resources are perfectly allocated, any intervention that creates change will necessarily create costs.  Theory admits for exceptions, and additional costs may be considered acceptable if the regulation solves a social problem.  However, the rule of thumb that ‘regulation increases costs’ and so should be avoided if at all possible has been repeated so often that it is ingrained in the mental models of many decision-makers, and even incorporated into institutional mandates.  The UK’s Better Regulation Executive, for example, has a mandate to ‘monitor the measurement of regulatory burdens and coordinate their reduction.’  A statement on climate change issued in 2019 by several thousand economists, including 28 Nobel Laureate economists and four former Chairs of the US Federal Reserve, urged governments to implement carbon pricing so that they could avoid the need for ‘cumbersome regulations’ that are ‘less efficient’.  

Steven Chu, Energy Secretary in the US under President Obama from 2009 to 2013, was so frustrated by the economists in his own department advising against the regulatory policies he knew were needed that he conducted his own study.  He found that contrary to the predictions of traditional theory, the imposition of energy efficiency regulations did not add to the costs of appliances such as refrigerators, washing machines and air conditioners; instead, by stimulating innovation, they accelerated those appliances’ reduction in cost.  A systematic review of relevant academic literature by Grubb et al found evidence of regulations being a major driver of innovation in lighting and building energy efficiency, and a significant driver of innovation in road transport.  

Disequilibrium theories of the economy help us understand why regulations can have such a positive effect.  In the constantly evolving ecosystem of a competitive market, laws or regulations set the rules of the game: they determine which technologies, products or strategies are the ‘fittest’, and which less so.  A regulatory change that introduces stringent new requirements immediately makes many products less fit for their environment than they were before, incentivising businesses to shift resources from exploitation (extracting value from current assets) to exploration (creating new assets).  The result is an acceleration of innovation, performance improvement, and cost reduction.

In situations where there is strong resistance to a low carbon transition, such that subsidies struggle to incentivise the desired change, regulations may be the most efficient of all policy options.  There is evidence to suggest that this is the case in the road transport transition, where taxes and subsidies have had relatively little impact, but regulations are proving highly effective at forcing car manufacturers to shift investment from petrol cars to zero emission vehicles.  The introduction of the EU’s latest regulations on 1 January 2020 saw electric vehicles’ share of car sales jump to 11% in that year, up from 3% in 2019.  California’s zero emission vehicles mandate has helped it achieve an electric vehicle share of car sales that is four times as high as that of the US as a whole.  The faster this shift of investment takes place, the more it strengthens the reinforcing feedbacks that improve the new technology, bring down its costs, and expand its market share.  

Rule 5: Tax should target tipping points 

The traditional rule of thumb for tax, or more specifically carbon pricing, is that it should be applied at a level that reflects the economic cost to society of each tonne of carbon.  In this way, the ‘externality’ of dangerous climate change is brought within the market.  A carbon price at this level is thought to maximise economic efficiency, by ensuring the optimum allocation of economic resources. 

There are two problems with this approach.  First, the total cost to society of climate change includes factors that are fundamentally uncertain, potentially catastrophic, and inherently subjective.  Consequently, there is no meaningful objective value for the ‘social cost of carbon.’   Second, in a low carbon transition – a context of disequilibrium – the aim of policy is not simply the efficient allocation of existing resources; it is the creation of new resources and new structures.  In other words, the aim is dynamic efficiency, not allocative efficiency.  In this context, the efficiency of a carbon price depends not on its absolute level, but on its relative level.  A carbon price that significantly alters the competitive balance between old and new technologies is likely to be more efficient than one that does not.  

When the problem is understood this way, we can imagine a new role for tax.  When clean technologies have been developed enough to begin to compete with fossil fuelled incumbents, tax – or tax together with subsidy – can tip the scales, helping the new to outcompete the old.  In dynamic systems, a tipping point is where a small intervention can lead to a disproportionately large change in behaviour.  This can happen because passing the tipping point activates new reinforcing feedbacks that create self-accelerating change.  In policy terms, this means getting a lot of bang for your buck.  

There is evidence that tipping points have played a role in the world’s fastest low carbon transitions in the power and road transport sectors.  In the UK’s power sector (decarbonising eight times faster than the global average), a fixed carbon tax made coal more expensive than gas at a time when both were being squeezed by the growth of renewables; as a result, coal crashed out of the system.  In Norway’s road transport sector, a combination of tax and subsidy makes electric vehicles cheaper at the point of purchase than equivalent petrol or diesel cars.  The share of electric vehicles in Norway’s car sales is around twenty times as high as the global average.  

Many more tipping points can be envisaged: such as when electric aeroplanes outcompete jet fuelled planes in short haul aviation; when hydrogen from renewables outcompetes hydrogen from gas; or when lab-grown meat outcompetes beef from cows.  Well targeted taxes can help us cross these tipping points more quickly, leading to rapid growth in clean technologies’ market share, faster performance improvement, and faster cost reduction.

Rule 6: When trading carbon, the narrower the scope, the better

Under Rule 1, we noted that a cap-and-trade scheme for carbon pricing creates a balancing feedback, which has a self-limiting effect.  In principle, this is undesirable when the aim is to stimulate rapid decarbonisation.  How much it matters in practice will depend partly on the scope of the scheme.  

A widely accepted rule of thumb about cap-and-trade schemes is that the broader their scope, the greater the efficiency.  This is derived from the presumption that a cap-and-trade scheme allows emissions cuts to be made wherever they can be made most cheaply.  The broader its scope – the more companies, sectors of the economy, or countries that are covered by the scheme – the more opportunities to cut emissions at low cost will be found, and the greater the overall economic efficiency will be.  

The mistake of this logic is to assume that we are dealing with an economy where nothing changes; where the objective is to discover the opportunities for least cost emissions cuts that have been left lying around.  On the contrary, our challenge is to create change in the economy, and this includes creating opportunities for low-cost emissions reduction that did not exist at the outset.  (Again, our aim is dynamic, not allocative, efficiency.)  There is no reason to assume that taking the least cost opportunity to reduce emissions at each moment in time will lead to the least cost transition over the course of time.  In fact, there is every reason to assume the opposite.  Early in a transition, the cheapest emissions reductions come from operating fossil fuelled technologies more efficiently (e.g. making a coal power plant more efficient, or switching from coal to gas).  This merely delays the investment that is needed to replace the entire stock of fossil fuelled capital assets with zero emission substitutes.  

If a cap-and-trade scheme covers a single sector in a single country, and has a steep trajectory forcing rapid emissions reductions, the opportunities to cut emissions in ways that are cheap but only delay the necessary transition may be exhausted relatively quickly.  The broader the scope of the scheme, the more opportunities there will be for the carbon price to be absorbed by (ultimately wasteful) marginal adjustments to the existing system, and the longer it will take for resources to be focused on creating and improving the new system. 

Rule 7: More is different 

Our decision-making processes sometimes assume that we can change one thing without changing anything else.  Cost benefit analysis, for example, is most appropriate for analysing the effect of a policy in situations ‘where the broader environment (e.g. the price of goods and services in the economy) can be assumed to be unchanged by the intervention’.  In such situations, we may be able to choose between policy options by considering them independently, one at a time.  

However, the aim of decarbonisation policy is not to leave the broader environment unchanged.  It is not merely to change the price of goods and services in the economy, but to create new goods and services, and new markets for them.  In such situations of disequilibrium, the behaviour of a system depends not only on the behaviour of its components, but on the interactions between them.  We will be best able to choose an effective set of policies if we consider them in combination, rather than individually. 

An example is provided by a disequilibrium modelling study of the road transport transition, by Lam and Mercure.  This simulation suggested that in China, the combination of a zero emission vehicle mandate, energy efficiency regulations, and a tax on petrol would achieve emissions reductions around 20% greater than the sum of the emissions reductions achieved by each of these three policies individually.  Meanwhile, other combinations of policies were found to yield an effect less than the sum of their parts.  If we remember that ‘more is different’, it may remind us to always look for policy combinations that achieve more than the sum of their parts, and not less.  

Rule 8: Sooner is better than later 

In an equilibrium economy, any change comes at a cost.  If decarbonisation is necessarily a net cost, then it makes sense to do the minimum required to meet our goals.  If our emissions need to follow a downward path, we may assume that we should make the smallest, easiest and cheapest cuts first, and leave the larger, more difficult and expensive changes as late as possible – especially if we assume that the economy grows (increasing our resources) and technologies improve over time.  

Equilibrium-based models have sometimes depicted decarbonisation as the task of Sisyphus.  They have assumed that any emissions reduction achieved one year needs to be paid for again the next, if it is to be repeated.  And they have assumed that the cost of reducing emissions at any time is independent of whatever has been done before.  So each year, policy pushes its boulder up the hill of decarbonisation, only to see it return endlessly to its starting point.  As larger emissions reductions are needed over time, the boulder is pushed higher each year, incurring ever greater costs that are only partially offset by technological improvement (assumed to happen by itself), and still it rolls back to its starting point.  Such would be our fate in an economy without structural change.   

Our experience of the transition to clean power already paints a different picture.  Solar and wind power are cheaper than coal or gas, becoming ever more so, and the transition to clean power can be made with a net economic benefit.  We can see that the same will be true of road transport, as electric vehicles move towards undercutting fossil fuelled cars.  Policies that put these zero emission technologies in place lead to emissions reductions that are permanent, not just temporary.  Disequilibrium modelling suggests that the entire transition to a zero emission economy could be made with a large net economic benefit, rather than a cost.   

This new picture is one in which the hill of decarbonisation is not endless: it has a summit.  In each emitting sector, we can push the boulder up the hill, starting with research and development, then with targeted investment and regulation, and maybe some tax near the top.  Then we can push it over the top and watch it accelerate down the other side, as the reinforcing feedbacks of the transition take over.  Sisyphus can be free at last.  

If the clean economy on the other side of the hill is more economically attractive than the fossil fuelled economy we are leaving behind, then the gains of this transition will be enjoyed in perpetuity.  There are many nuances, but to a first approximation, the sooner we are able to start enjoying these gains, the better.  Faster deployment of clean technologies leads to faster cost reduction, and earlier arrival at the point where the transition begins to yield net benefits.  For this reason, a recent study finds a fast global transition would cost less (and achieve greater net gains) than a slow transition.  

Rule 9: Assess opportunities and risks, not just costs and benefits 

We are likely to make good policy decisions more often if we use decision-making frameworks that are appropriate to the context.  Cost benefit analysis can be useful in the special circumstances when the costs and (future) benefits of a policy can be predicted and quantified with reasonable confidence, and when any changes provoked in the economy are expected to be marginal (having no effect on the economy’s structure).

Decarbonisation policy takes place in a different, and more general, context.  Many important effects of policies – such as their ability to create new technologies and markets – cannot be predicted quantifiably with reasonable confidence.  Limiting analysis to quantifiable factors would therefore be misleading; it is essential to broaden the scope of analysis from costs and benefits to risks and opportunities.  Since the intent of policy is to create transformational (not marginal) change, this ‘risk opportunity analysis’ must compare policy options in terms of their effects on processes of change within the economy, not simply in terms of expected outcomes at a moment in time.  This analysis can be informed by models and theory that simulate and explain the behaviour of the economy in any of its possible dynamic states, not limited to the special (and in this context inapplicable) case of equilibrium.

With some help from hindsight, it is possible to see how the application of a ‘risk opportunity analysis’ approach could have led to recommendations in favour of the decarbonisation policies that turned out to be so successful in China, India, Brazil and the UK, where cost benefit analysis had generally advised against.  

Rule 10: Work together to make progress faster 

When decarbonisation was assumed to come at a net cost, the diplomacy of climate change could only be a negative sum game.  Much effort was expended seeking to agree a division of the finite global carbon budget, with countries individually aiming to maximise their share of the remaining ‘carbon space’.  As a global agreement along these lines proved impossible, the attempt was abandoned after the Copenhagen climate change conference of 2009.  The Paris Agreement restored confidence in the collective effort by allowing each country to set its emissions targets unilaterally.  

The new understanding that decarbonisation can bring a net economic gain creates the potential for a new kind of climate change diplomacy.  Cooperation can be positive sum.  Coordinated international action can bring faster innovation, stronger incentives for investment, larger economies of scale, and level playing fields where they are needed, making decarbonisation faster, easier, lower cost and greater gain for all countries.  Since each of the emitting sectors is different in its political, financial, technological and industrial structures, most of these coordination gains can only be accessed through targeted cooperation in each sector.  

This was the vision of the leaders of more than 40 countries, covering over 70% of global GDP, that at the COP26 climate change talks in Glasgow signed up to the Breakthrough Agenda – committing to work together to make clean technologies and sustainable solutions the most affordable, accessible and attractive option in each emitting sector before the end of this decade.  Creating and strengthening institutions to support the strong, targeted and sustained cooperation that is needed in each sector must now be a high priority for the international community.  

Keeping our hopes alive 

It may be too late to achieve the international community’s original goal of ‘preventing dangerous anthropogenic interference with the climate system’.  We have already interfered, and it has already proved dangerous.  But perhaps the 1.5 degree goal, while outrageously difficult given our slow start, is not altogether impossible.  The great variation in the pace of transition between countries in a given sector suggests that if more countries put the most effective policies in place, a much faster global transition is possible.  The potential gains from cooperation in each sector, so far relatively unexplored, could yield a significant further acceleration.  

The assumption underlying the approach proposed here is not that we have any greater political will for achieving decarbonisation.  It is only that we are dealing with the economy in a state of change, rather than a state of equilibrium.  This understanding can guide us to better choices, so that the same political will and financial resources can be deployed to considerably greater effect.