Dimitri Zenghelis explains why Paul Romer’s theory of endogenous growth can be harnessed to direct and design a net-zero-carbon future while using William Nordhaus’s DICE and RICE models may already have proved truly damaging.
Last year William Nordhaus and Paul Romer jointly received the Nobel Prize in Economic Sciences: Nordhaus for his work on the damage caused by climate change and Romer for developing endogenous growth theory, which examined how economies can achieve a healthy rate of economic growth.
However, there is reason to believe that Nordhaus’s work has actually slowed humanity’s progress in addressing climate change by unnecessarily discouraging action. By contrast, Romer’s theory, which has mostly been ignored in the field of decarbonisation, gives us a full understanding of exactly what to do to address climate change. So exactly how do the two approaches differ in assessing humanity’s ability to decarbonise our economies?
Nordhaus’s models: bad forecasts become self-fulfilling prophecies
Nordhaus constructed integrated economic and scientific models (the DICE and RICE models) to determine the efficient path for coping with climate change. These models are widely used today by policymakers. The models presuppose the technologies, tastes, preferences and behaviours that will dominate in the decades and centuries ahead.
But these imposed structural assumptions are precisely the things we need to know when predicting the costs of decarbonisation. How will technologies emerge, how will tastes and preferences change, what underlies a shift from one energy and production network to another, how do consumers change habits and social norms? These things will determine the ease with which we might combat climate change. In many cases, these elements are subject to great inertia until they reach a tipping point where expectations change rapidly and technologies switch from one network to another.
For example, as enough players shift investment and new technologies are deployed, learning and experience improve performance and lower the costs of clean technologies. The cost of finance in what were formerly considered niche markets would also fall. The development of new business lobbies, supportive institutions and behaviours and new skills would be expected to reduce unit costs further. This makes deployment of these technologies even more attractive until incumbent technologies, products and networks become obsolete. Those late to recognise the transition stand exposed to stranded or devalued assets. Such structural shifts are a key feature of innovation and adoption.
For years economists have talked up the costs of decarbonising energy generation. Yet costs have collapsed. None predicted the precipitous fall in the price of renewable technologies. Solar photovoltaic (PV) costs fell 44 per cent in the two years to the end of August 2017 and have fallen by 83 per cent since 2010, a period over which the price of wind turbines has dropped 35 per cent. None predicted that renewables would become the dominant source of energy investment in the first quarter of this century – or the expansion in LED lighting from less than 5 per cent to more than 40 per cent of the global market in the past six years. Yet followers of Paul Romer would have known that these processes are predictable.
When it comes to the damages from climate change, models like RICE and DICE are particularly ineffective at providing forecasts because they translate the uncertainties in the science poorly. MIT economist Robert Pindyck claims models tell us “very little” about climate change and how we solve it. Nobel laureate Thomas Schelling goes further. Even if we had perfect information on the climate system for the next century, to pretend that we know anything about the economic system that will exist at that point is absurd. Imagine if you were tasked with that question in 1900, for 2000. You would consider the impact on health, but for a world not only without MRI scanners but also without widespread handwashing (Typhoid Mary was just getting started in 1900).
But the problem is worse than one of spurious precision and bad forecasts. Nordhaus and those who use static models not only get the future wrong, they make the future wrong by generating what game theorists call an inferior Nash equilibrium. To the extent that such models are believed, they become self-fulfilling. Models that overstate the cost of decarbonisation discourage businesses and policymakers from investing in renewables and energy efficiency. This slows the rate at which the costs come down, making decarbonisation harder.
Incorporating these features of path-dependent phenomena – knowledge spillovers, network effects, switching costs, feedbacks, and complementarities as well as irreversible and potentially catastrophic climate impacts – into economic models is difficult and often leads to a multiplicity of ‘equilibria’, each dependent on a different development path. Yet they are features of the real world.
Romer’s theory is based on steering growth and investing in the future
Romer understood all these dynamics and feedbacks and, by contrast, concentrated on how expectations and actions determine outcomes. His theory of endogenous growth, for which the Nobel was awarded, is based on the insight that growth in total factor productivity and innovation do not just happen. They need to be steered. Growth, he showed, is the result of investments made in human, physical and other forms of capital. Growing knowledge associated with these investments generates complementarities and networks.
For example, investing in computers induces smart ideas about how to use them. The ideas and associated software increase the returns to investing in computers and broader ICT networks. Investment in physical and knowledge capital begets increased output and generates resources for further investment. In this way, increasing returns to ideas overcome the diminishing returns to factors like labour and capital.
What has this got to do with addressing climate change? Well, it means that both the speed and nature of growth, for example whether it is clean and sustainable or dirty and based on resource depletion, will depend on the policy choices undertaken today and the infrastructure, technologies and institutions we seek to lock in to. It will also depend on shaping expectations. The cost of decarbonisation in decades to come will be a function of the action and investment taken today. The correct answer to the question ‘what will it cost?’ cannot be delivered by a static model because it depends on what is done now. This is an important policy message.
To illustrate the point, Romer makes the important distinction between ‘complacent optimism’ and ‘conditional optimism’. He writes: “Complacent optimism is the feeling of a child waiting for presents. Conditional optimism is the feeling of a child who is thinking about building a treehouse. ‘If I get some wood and nails and persuade some other kids to help do the work, we can end up with something really cool.’” He adds: “What the theory of endogenous technological progress supports is conditional optimism, not complacent optimism. Instead of suggesting that we can relax because policy choices don’t matter, it suggests to the contrary that policy choices are even more important than traditional theory suggests.”
Ignoring political, cultural and institutional barriers means being wrong about the future
What Romer is saying is that the barriers to overcoming major challenges are not primarily economic or technological, they are political, cultural and institutional. Models that fail to understand this will consistently get the future wrong, as I argued in a recent lecture series.
What undermines Nordhaus’s approach is that we cannot accurately predict the technologies, behaviours or institutions that will deliver a net-zero-carbon world, nor do we know the precise costs. It is foolish and counterproductive to pretend we do. Counterproductive because, as Romer points out, one of the things we do know is that these costs will be a direct function of the decisions, choices and investment decisions we make today. We also know that after periods of inertia change can come fast and, once it comes, can be transformative, with new networks, behaviours and institutions replacing old.
The bottom line, then, is that when faced with systemic technological transformation, economists, policymakers and investors should spend less time using models like Nordhaus’s to fruitlessly attempt to predict the future and more time using models like Romer’s to direct and design it. Romer was right. The cost of addressing climate change is endogenous.
This article first appeared on the blog of the London School of Economics.