How Do We Think About and Respond to Climate Change?

Class Essay

I'm currently taking ECON 309: Economic Problems and Public Policies with Professor Chris Coyne at George Mason University. I would highly recommend both the class and the professor. For our final project, we were tasked with having a conversation with AI about a public policy issue of our choice and then writing an essay based on that discussion. This is the essay I wrote. I hope you enjoy it, and I’d welcome any suggestions before I submit it.

This essay examines climate change through the economic way of thinking, drawing from Chapter 25 of the textbook, outside knowledge, and insights from my conversation with AI.

My Views On Climate Change Before the AI Dialogue

In my view, climate change represents a market failure—a situation where individuals pursuing their self-interest in a free market fail to produce the socially optimal outcome. Specifically, climate change is a negative externality. Negative externalities are costs imposed on third parties that aren't reflected in market prices. In the case of climate change, greenhouse gas emissions harm society at large, yet emitters bear only a fraction of the total costs. As a result, neoclassical economic theory predicts that free markets will generate more pollution than is socially optimal (where the marginal benefits of carbon pollution equal the marginal costs).

Economists identify three ways to correct this market failure: governments can directly limit emissions through command-and-control regulations; they can establish a cap on emissions and allow firms to trade permits; or they can impose Pigouvian taxes that price emissions to close the gap between private and social costs. Most economic textbooks argue that the latter two methods are the most cost-effective ways to achieve climate mitigation.

In my paper for the George Washington University Undergraduate Economic Review, I demonstrate that permits are the most cost-effective way to reduce pollution in theory, but Pigouvian taxes are usually the most cost-effective way to reduce pollution in practice. If we assume perfect implementation, tradable permits are the most effective way to reduce pollution because the burden of emissions reductions is borne by those firms or countries with the lowest marginal abatement costs. However, examining the European Emissions Trading System reveals how energy producers lobbied for grandfathering clauses and an artificially cheap price for permits. Thus, the initial impact of the program was to increase energy costs without a comparable reduction in carbon emissions. By contrast, carbon taxes are less prone to rent-seeking or implementation errors and reduce pollution more cost-effectively than either imperfect implementation of permits or traditional regulation.

Thus, like most economists, I favored a carbon tax. However, an experience I had at the Cato Institute laid the seeds of doubt about the desirability of a carbon tax, and these seeds have only ripened from my conversation with AI.

Reflections on AI Dialogue

I started my conversation with AI by pointing out a disconnect. Most economists advocate for a carbon tax as the most effective way to address climate change, yet the United States relies on regulation and subsidies instead. At the same time, forty-six countries, including Sweden, the United Kingdom, and Canada, use some form of carbon pricing. The AI and I explored several reasons why the United States has not adopted a similar approach.

First, the U.S. fossil fuel sector is massive and politically well-organized, especially in swing states like Pennsylvania. Second, Americans have long been averse to taxes (hence the slogan "No taxation without representation"). Third, carbon taxes have visible costs—you have to pay more at the pump, for instance—but subsidies are invisible. You don't know how much of your tax dollars go toward subsidizing oil, wind, solar, and so on. Fourth, Americans are reluctant to take action on climate change because they see countries like China and India dramatically increasing their carbon emissions. This is partially a matter of fairness, but also a concern about free riding—if the United States cuts CO₂ emissions, other countries can enjoy the benefits while paying none of the costs.

However, there are also economic problems with the carbon tax. Allow me to digress with a brief anecdote if I may. I interned at the Cato Institute this past semester, and we attended a series of seminars, including one by their scholar on energy and environmental policy, Travis Fisher. Fisher gave a presentation on energy abundance and at the end of the presentation I asked the following: "If climate change imposes negative externalities and we can't know the most cost-effective way to reduce emissions because of the knowledge problem, wouldn't it be better to scrap all our energy subsidies and impose a carbon tax equal to the size of the externality?" The other scholars in the room laughed, and I was somewhat confused. Fisher explained that in the past years, he had given his whole one-hour presentation on the problems with the carbon tax, and he had to condense it into five minutes to answer my question. Fisher went on to say that economists have a wide range of estimates about the social costs of carbon. Moreover, the production of almost every good and service generates carbon emissions at some point in the production process. As a result, a carbon tax would give the government the authority to impose an arbitrary tax that hits every sector of the economy. He then quoted P.J. O'Rourke that giving the government money and power is like giving a teenage boy car keys and whiskey. I sat on Fisher's comments about the carbon tax during lunch and somewhat melodramatically claimed I was having an existential crisis. "How could my first published economic paper be wrong?" "Why can't carbon taxes work? The graph in Econ 101 is so beautiful, as if written by the hand of divinity himself."

Eventually, I concluded that the carbon tax had serious problems, but I didn't know how to fix those problems or if a better alternative existed. I also reflected that in Econ 101, my professor put heavy emphasis on market failures. Fair enough, market failures exist and are worth studying. However, after talking about market failure, my Econ 101 professor went on to describe a perfect government on a graph with far fewer complexities than the world we inhabit. I realized that we need to compare imperfect markets with imperfect government and then decide which is the most tolerable of imperfections.

After I told the AI about my reservations and problems with the carbon tax, the AI stated that most economists realize that a carbon tax is only one part of the cocktail to address climate change. I added that a neglected component of this cocktail was adaptation. As Bjørn Lomborg explains in his book False Alarm, many estimates about the damage of climate change assume static human action and thus overstate the true costs of climate change. In reality, we have already seen humans adapt to the impact of climate change. Climate change will increase sea levels, heat waves, and the frequency and severity of natural disasters. However, even in a world with climate change, the number of climate-related deaths has decreased by 95% since 1950. This is because humans have built better dams, more air conditioning, and more resilient buildings.

The AI and I also discussed how humans are often bad at predicting the future and evaluating risks. Humans overstate some risks (e.g., overpopulation, resource depletion) and underestimate other risks (e.g., financial crises and pandemics). Maybe we should be worried about asteroids crashing into Earth. We concluded that the best way to balance known and unknown risks is through wealth and technology, which make it easier to respond to a variety of problems.

I asked the AI why so many environmentalists were hostile towards economic growth and capitalism. The AI pointed out that environmentalism (Paul Ehrlich's Population Bomb, the Limits to Growth report) has historically emphasized the limits of population and economic growth. I would like to point out that these predictions have not come to pass. The Stanford biologist Paul Ehrlich's 1968 book The Population Bomb predicted mass starvation and resource wars by the 1970s and 1980s, yet global food production has increased faster than population growth. The 1972 Limits to Growth report, based on models developed at MIT, warned that the world would run out of key resources, such as oil, copper, and gold, within a few decades. However, technological advances and new discoveries have continually expanded available reserves. Both works underestimated human ingenuity, technological progress, and market mechanisms to overcome constraints.

However, one concern I expressed is that the rate of economic growth in rich countries has slowed down since the 1970s. This makes it more difficult to expand the economy's production possibilities frontier and respond to new problems. One explanation for why economic growth has slowed is the Robert Gordon-Tyler Cowen argument that we have entered a period of technological stagnation. Gordon contends that the transformative innovations of the late 19th and early 20th centuries—electricity, internal combustion engines, modern sanitation, and telecommunications—created a one-time boost to productivity that cannot be replicated. Meanwhile, Cowen argues that we are experiencing "The Great Stagnation," where low-hanging fruit in terms of technological advancement, education, and institutional improvements has been largely exhausted. Both scholars suggest that recent digital innovations, while impressive, have had relatively modest impacts on broad-based productivity growth compared to earlier breakthrough technologies. This technological plateau, combined with demographic headwinds, helps explain why developed economies struggle to achieve the robust growth rates of previous decades.

If Gordon and Cowen are correct, as the AI points out, this makes both climate mitigation and adaptation more costly. Humanity would be less able to transition to green energy, which heavily relies on technological breakthroughs in areas such as energy storage, fusion power, and affordable carbon capture. These breakthroughs could face substantial delays if innovation stagnates. In the meantime, climate risks may begin to outpace our innovative capacity, as a stagnating economy would have fewer resources available for critical resilience measures, infrastructure upgrades, and disaster response capabilities. This challenge would be particularly acute for developing nations, as global growth stagnation could severely impact the ability of the rich world to provide development aid, maintain robust trade relationships, and facilitate the diffusion of technology to poorer regions.

Thus, while there are no free lunches when it comes to climate change or any other issue, increasing the long-run sustainable growth rate would give us cheaper and tastier lunches to choose from.

Conclusion

Overall, I was quite impressed with AI's capabilities to engage in substantive discourse. My point of comparison is the types of debates I have with my friends about economics and politics, often over copious amounts of alcohol. On the one hand, I would have appreciated it if AI had given more pushback and told me when it thought I was spouting nonsense, which would have made our interaction feel more human. On the other hand, my conversation with AI helped me organize and articulate my thoughts on climate change more effectively than I would have been able to in a live human interaction.

I was able to explore the potential and drawbacks of carbon taxes and adaptation through economic growth. The synthesis I reached from my questions to AI is that there will always be tradeoffs in a world of political and economic constraints. However, economic growth can give us better options to choose from.

Admittedly, my ecomodernist view may be overly optimistic and outdated. AI raised several compelling arguments for why climate change could be different from past environmental challenges. For example, climate change poses risks to ecosystems that may be less able to adapt to environmental changes than humans, and climate change may reach a tipping point where irreversible changes occur rapidly and simultaneously across multiple systems. It is hard to evaluate these claims as someone with no background or expertise in the hard sciences.

Nevertheless, my conversation with AI underscored that the economic way of thinking is necessary but not sufficient for analyzing complex problems.