Model reveals how disinformation can trigger abrupt collapse of carbon taxes

Carbon pricing has long been treated by economists as the single most powerful lever against climate change, yet the carbon taxes actually implemented around the world remain stubbornly far below the levels suggested by benchmark climate-economy models. A new study argues that the gap may have less to do with the physics of the climate system or the mathematics of optimal taxation than with something far more fragile: what voters believe about climate damages. When disinformation distorts those beliefs, the political foundations of ambitious carbon pricing can erode, and the new modeling shows that the erosion may not be gradual at all. Instead, policy can slide off a cliff.

The research, published in the KeAi journal Risk Sciences by Katsiaryna Bahamazava of University College Dublin and Stanley Reznik of the ILaVita Foundation, breaks with a standard assumption that has quietly underpinned decades of integrated assessment modeling. Most climate-economy models treat citizens as holders of accurate beliefs about the damages warming will inflict. In reality, public understanding is shaped by learning, by organized disinformation campaigns, and by the visible consequences of climate change itself. The authors built these dynamics directly into a dynamic climate-economy model, dividing the population into two groups: informed agents who correctly perceive the true extent of climate damage, and misinformed agents who systematically understate it. The share of informed agents evolves over time as people learn, as disinformation shocks hit, and as observable climate harm feeds back into beliefs.

Within this framework, the researchers compared two distinct mechanisms for setting a carbon tax. The first is a politically constrained planner, a decision-maker who understands the true damages but faces implementation costs whenever policy ambitions exceed what public opinion will actually tolerate. The second is majority voting, in which policy is chosen by the median voter and adjusts only gradually, reflecting the well-documented sluggishness of real-world tax systems. This setup allowed the team to ask a question that conventional models, by assuming perfect information, simply cannot pose: does democracy help or hurt climate policy when the electorate’s beliefs can be manipulated?

The answer, at least initially, favors democracy. Contrary to the widespread view that insulated planners are always better suited to long-horizon climate policy, the model shows that majority voting can generate stronger carbon taxation than a politically constrained planner. As long as informed voters remain the majority, the median voter backs an ambitious policy, producing stronger early mitigation and lower peak warming. The intuition is that a planner constrained by public opinion must constantly hedge, tempering policy to avoid implementation costs, whereas an informed electorate can sustain a more aggressive tax because the political constraint and the policy preference coincide. The authors describe this as a democratic advantage: informed majorities deliver more ambitious mitigation than a planner who must negotiate with a potentially wavering public.

That advantage, however, turns out to be startlingly fragile. Once disinformation erodes the informed majority and misinformed voters take over, the political target for carbon taxation shifts sharply downward, and implemented policy declines rapidly. The authors call this a cliff-edge dynamic, and the metaphor is apt: rather than a smooth tapering of ambition, the model produces an abrupt drop in taxation once the median voter flips. The politically constrained planner, by contrast, weakens more smoothly as public support fades, adjusting policy in step with the gradual erosion of the opinion constraint. The asymmetry matters for policy design, because it suggests that democratic climate policy is not merely vulnerable to disinformation but vulnerable in a discontinuous, hard-to-reverse way. Small shifts in the belief distribution may have little visible effect on policy until a threshold is crossed, at which point the political equilibrium collapses.

Timing emerges as a second critical dimension. Disinformation arriving late in the transition, or persisting over an extended period, proved especially damaging. The reason is structural: late or persistent disinformation strikes after the shift to stronger climate policy has already begun, undermining the political coalition just as it is needed to sustain the higher tax path. In the model, these scenarios generate what the authors call hothouse warming trajectories under both political regimes, with peak temperatures and final atmospheric carbon dioxide concentrations substantially worse than in the benchmark case. Early disinformation, by contrast, can be partially absorbed by subsequent learning and by the feedback from visible climate harm, which gradually restores the informed share of the population. The practical implication is uncomfortable: the period when climate policy is gaining momentum may be precisely the period when it is most exposed to belief manipulation.

The study is not purely a warning. The researchers also modeled counter-disinformation interventions, and the results offer a measure of encouragement. A sufficiently strong intervention, one that counteracts the disinformation shock and restores the informed share of the population, almost completely offset an early disinformation shock, returning climate and policy outcomes close to the no-disinformation benchmark. In other words, the cliff-edge dynamic is not irreversible if the information environment is repaired in time. Peak warming, final atmospheric carbon dioxide, early carbon taxation, and average abatement all recovered toward benchmark levels in the intervention scenario, suggesting that investments in information integrity can function as genuine climate policy rather than as a separate concern about democratic quality.

This framing is one of the study’s most provocative contributions. Protecting the information environment, the authors argue, is not merely a matter of public discourse or democratic health; it is a macroeconomic component of effective climate policy. Information policy is not just complementary to carbon pricing, they contend; it helps sustain the political conditions under which carbon pricing remains feasible at all. In a model where the tax path depends on the median voter’s beliefs, spending on countering disinformation becomes, in effect, an investment in the durability of the carbon price itself. That reframing has concrete implications for how governments budget for climate strategy, potentially placing communication and information-integrity programs on the same analytical footing as technology subsidies or infrastructure investment.

The methodological approach is worth emphasizing. The study is a computational simulation, not an empirical measurement of any particular country’s politics, and its strength lies in its ability to compare counterfactual scenarios, benchmark conditions without disinformation, early shocks, late shocks, persistent disinformation, and intervention cases, within a single consistent framework. By endogenizing beliefs, allowing them to shift through learning, disinformation shocks, and feedback from realized climate damage, the model captures a feedback loop that standard integrated assessment models omit entirely: climate policy depends on beliefs, beliefs depend on climate outcomes and on the information environment, and the information environment itself can be strategically manipulated. The comparison between the constrained planner and majority voting further isolates how political institutions mediate the damage that distorted beliefs can do.

For a policy community increasingly alarmed by the spread of climate misinformation, the study offers both a conceptual tool and a caution. The conceptual tool is the recognition that the relationship between public belief and carbon pricing is nonlinear, with democratic majorities capable of outperforming constrained planners until a tipping point is reached. The caution is that the most dangerous disinformation may be the kind that arrives late, when policy momentum is highest and the political coalition supporting it is most needed. The model’s hothouse trajectories under late and persistent disinformation scenarios quantify what is at stake: materially higher peak warming and atmospheric carbon concentrations under both political regimes. If the authors are right, then the fight against climate disinformation is not a sideshow to carbon taxation. It is, in a very literal sense, one of the mechanisms that determines whether the carbon tax survives.

Subject of Research: Modeling the effects of climate disinformation on the political feasibility and timing of carbon taxes

Article Title: Climate disinformation can push carbon taxes off a cliff model shows

Article References: Climate disinformation can push carbon taxes off a cliff model shows. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: carbon tax, climate disinformation, climate-economy modeling, majority voting, median voter, political economy, belief dynamics, counter-disinformation, peak warming, Risk Sciences, computational simulation, climate policy