The cost of climate mitigation revisited

I know, we have bigger fish to fry in the age of Covid-19, but one of the things we need to do as well, is think about how to kickstart our economy after the pandemic is over and how to put it on a more sustainable path. Climate change may get some respite from the lower emissions we produce this year, but it won’t go away.

And here is my beef with most climate models: they tend to assume that mitigating climate change and reducing greenhouse gas emissions is only a cost and has no benefits. Furthermore, most climate models assume that the cost of reducing greenhouse gases is constant and does not decline as we develop new technologies.

There certainly is a cost element to climate mitigation insofar as it requires us to reduce the consumption of some goods and services. In the usual GDP accounting, this reduced consumption leads to lower GDP growth.

However, climate mitigation also requires us to invest in new technologies and these investments lead to higher GDP growth. Furthermore, consumption might not just be reduced in one area but also increased in other areas. Switching from cars powered by petrol to electric cars certainly reduces the consumption of petrol cars, but it also increases the consumption of electric cars. It is not clear that this switch from one technology to another leads to higher costs and lower GDP growth in the long run.

While electric cars are more expensive than petrol cars today, their costs are declining rapidly as the new technologies mature. Current projections show that in 2025, electric cars will be cost-competitive with petrol and diesel cars and from then on, buying an electric car will be cheaper than a car with an internal combustion engine. Yet, the models of climate mitigation assume that reducing greenhouse gas emissions costs the same in 2050 than it does today as Michael Grubb and Claudia Wieners point out in a blog post for the Institute for New Economic Thinking. In other words, these models assume that buying an electric car will always be more expensive than buying a petrol or diesel car.

Under these assumptions of time independence of the costs of climate mitigation, the sensible way to deal with climate change is to wait a long time to initiate action until national income has grown large enough to make climate mitigation affordable compared to the imminent costs to the economy from climate disasters. Hence, these models recommend limited climate mitigation and then an inevitable policy response in the face of impending disaster that leads to a revolution in our economy and a quick, disruptive transition.

But the experience of the ongoing transition from fossil fuels to wind and solar energy shows that costs of climate mitigation do not stay constant over time. Instead, they are falling rapidly and the more wind and solar power stations are installed today, the cheaper they will become in the future, simply because businesses are getting better at building windmills and solar cells. Wind and solar energy are already the cheapest form of energy in almost every country today (even without subsidies) and far cheaper than coal, for example. It is uneconomical to keep existing coal power plants online in many countries and in a few years’ time, it will be cheaper to build new solar or wind power plants and actively shut down existing gas power plants.

Grubb and Wieners explicitly model the cost of climate mitigation in a world where the costs for switching from old to new technologies depend on past mitigation efforts. I do not want to bore you too much with the maths but the key variable is the abatement parameter p. If p = 0 future climate mitigation costs do not depend on past mitigation efforts at all. If p = 1 mitigation costs depend only on the rate of change, i.e. the increase in mitigation efforts relative to the past (i.e. businesses have factories in place and they only have to invest in expanding their factories to build more windmills, etc. but have no additional costs to keep production at past levels). For a value of p = 0.5, abatement costs are a 50/50 mix between these two extremes. There are arguments to be made that p should be close to 1 and Grubb and Wieners even try to argue that in an environment where businesses learn from experience, p could be larger than 1. In this case, there would be economies of scale to climate mitigation and abatement costs would decline faster and faster the more experience we get with new technologies.

In any case, here are the estimated costs to GDP from climate mitigation efforts in this more realistic model. If mitigation costs remain independent of past mitigation efforts (p = 0) the costs of climate mitigation to GDP would increase steadily because we are emitting more greenhouse gases every year, making the task to clean up our atmosphere bigger every year. If, on the other hand, mitigation costs only depend on how much we do relative to the past (p = 1) the costs on GDP would be relatively high in the next decade or so, but then decline quickly in the 2030s and 2040s as the global economy has switched to technologies that don’t harm the climate. It is obvious from the chart below that the cost of climate mitigation is not zero in any case, but it can likely be minimised by transforming our economy sooner and gradually rather than later and suddenly.

As we move past the current crisis, the economy will likely need additional government spending to catch up with lost output. What better than to invest in renewable energy infrastructure and other measures to mitigate and adapt to climate change?

Source: Grubb and Wieners (2020).