Introduction

The costs of unchecked climate change pose devastating risks to the global economy, potentially causing significant reductions in global income and worsening global income inequality. In the United States, the poorest counties are projected to face damages as high as 20% of county income by the late 21st century without mitigation efforts. Despite these net benefits, broad-based political consensus for large-scale mitigation remains elusive and increasingly polarised. This research presents a novel approach for building support for climate action by giving individuals financial exposure to energy sector stocks impacted by the green transition. The study argues that ownership and trading opportunities encourage individuals to learn independently about climate change risks and the returns of mitigation policies.

Research design

The study utilised a large-scale randomised controlled trial (RCT) involving 3,806 American adults with diverse socioeconomic and political backgrounds. Participants were recruited from across the United States, with over-sampling in states particularly exposed to climate risks, including Alabama, Georgia, Kentucky, Mississippi, North Carolina, South Carolina, Ohio, Pennsylvania, Tennessee, West Virginia, Arizona, and Louisiana. The experiment randomised financial market exposure to the energy sector and the specific type of portfolio received: either fossil-fuel (brown) or renewable energy (green) assets.

Treated respondents received initial portfolios with values of $50, $100, or a ‘fantasy’ allocation of $100 with no real monetary value. Over a six-week period, participants could track and trade their assets on a curated platform. The study employed four survey waves: baseline, midline (three weeks post-treatment), endline (seven weeks post-treatment), and a long-term follow-up eight months after treatment. To prevent social desirability bias, the climate attitude surveys were administered anonymously and separately from the trading module.

Empirical strategy

The primary specification used a pre-registered OLS regression to estimate the Intent-to-Treat (ITT) effect of exposure to financial markets. All specifications included LASSO-selected baseline covariates and fixed effects for randomisation strata. The analysis controlled for climate disclosure status in endline and long-term surveys, as some participants received specific environmental disclosures during the trading module.

Main outcomes overview

Support for climate action was measured using a composite z-score called the Climate Action Support Index (CASI), comprising nine sub-indices. These sub-indices captured climate beliefs, perceptions of transition benefits and tradeoffs, policy preferences for government and business action, and pro-environmental behaviours. Behavioural measures included an incentivised willingness-to-pay measure through donations to the Environmental Defense Fund and self-reported intentions regarding longer-run life decisions, such as where to work or live.

Results

The pooled financial exposure treatment raised the overall CASI by 0.036 standard deviations (SD) at midline (p-value 0.015), increasing to 0.065 SD by the endline (p-value < 0.001). The effects were significantly stronger for those with real assets (0.072 SD) compared to fantasy treatments (0.035 SD). Ownership of both green and brown assets caused positive support for mitigation, with those assigned green assets showing an effect of 0.089 SD and brown assets 0.057 SD. The results were most pronounced among individuals who were initially more climate-skeptic.

Learning mechanisms

The study provides evidence that learning is the primary driver of these shifts. Treated respondents substituted away from polarised media sources towards financial news outlets, such as the Wall Street Journal and Yahoo Finance, to understand the value of their assets. This pattern suggests that financial exposure incentivises personal learning about fundamentals and market implications.

Media consumption

Exposure to financial markets increased the consumption of financial news by 0.088 SDs at midline and 0.102 SDs at endline (p-value < 0.001). Conversely, treated participants were significantly less likely to obtain news from social media (-0.139 SDs) or Fox News (-0.080 SDs) at endline. This indicates a shift in information diets towards sources providing relevant financial data over polarised content.

Knowledge and perception of environmental impacts of energy assets allocated in the treatment

Treated respondents demonstrated greater objective knowledge of firm-level greenhouse gas emissions, identifying emission levels 0.160 SDs more accurately than the control group. Subjective evaluations also shifted; treatment increased the perceived environmental impact of brown firms and reduced it for green firms. Participants were able to distinguish between assets in different phases of the green transition more clearly after treatment.

Financial knowledge and confidence

By the endline, respondents gained broader financial knowledge and confidence. The treatment effect for the financial confidence index was an increase of 0.067 SDs (p-value < 0.05). This effect was primarily driven by those allocated real assets, with an estimated effect of 0.088 SDs (p-value < 0.01). Knowledge gains were similar regardless of whether participants held green or brown portfolios.

Long term effects

A final follow-up survey eight months post-treatment confirmed that effects persisted after the assets were divested. The overall pro-climate preference remained positive at 0.038 SDs (p < 0.05). Treated respondents continued to hold more positive beliefs regarding the benefits of the green transition and maintained higher levels of support for government-led mitigation action (0.087 SDs, p < 0.05).

Additional analyses and robustness checks

Heterogeneity analysis revealed that treatment effects were most significant for individuals with below-median support for climate action at baseline. For this subgroup, effects were four times larger at midline and five times larger eight months post-treatment compared to the full sample. Robustness checks using alternative covariate selections and accounting for attrition showed consistent results.

Conclusion

The research demonstrates that exposure to financial markets can foster durable support for climate mitigation. By incentivising individuals to seek credible information independently, the intervention sidesteps polarised debates and allows participants to reach their own conclusions. These findings highlight the potential of financial innovations, such as passive investment vehicles in pension plans, to increase awareness of climate stakes and encourage pro-environmental policy support.