Amplified challenges for the joint reduction of emissions and inequality, considering the missing-rich

Existing studies find that reducing income inequality increases greenhouse gas emissions. However, these analyses usually rely on household surveys that underrepresent top-income groups or on corrections that assume a constant relationship between income and emissions, both of which can distort the results. This study develops a new approach that combines tax microdata to correct the income distribution with machine-learning-based predictions of consumption to better estimate emissions among high-income households. Applied to Ecuador, a highly unequal country with institutional commitments to environmental protection, the approach identifies substantially higher emissions at the top of the income distribution and a non-monotonic pattern of emissions intensity per US$ of income that conventional methods fail to detect. Simulations show that full redistribution raises emissions by 26% under the new approach, compared with 6% in uncorrected surveys. These results underline the importance of better data and methods for evaluating inequality-emission trade-offs and designing redistribution policies.