Abstract: Sustainability policies are often motivated by the potential to achieve multiple goals, such as simultaneously mitigating the climate change and air quality impacts of energy use. Ex ante analysis is used prospectively to inform policy decisions by estimating a policy’s impact on multiple objectives. In contrast, ex post analysis of impacts that may have multiple causes can retrospectively evaluate the effectiveness of policies. Ex ante analyses are rarely compared with ex post evaluations of the same policy. These comparisons can assess the realism of assumptions in ex ante methods and reveal opportunities for improving prospective analyses.

We illustrate the benefits of such a comparison by examining a case of two energy policies in China. Using ex post analysis, we estimate the impacts of two policies, one that targets energy intensity and another that imposes quantitative targets on SO₂ emissions, on energy use and pollution outcomes in two major energy-intensive industrial sectors (cement, iron and steel) in China.

We find that the ex post effects of the energy intensity policy on both energy and pollution outcomes are very limited on average, while the effects of the SO₂ emissions policy are large. Compared with ex ante analysis, ex post estimates of benefits of the energy intensity policy are on average smaller, and differ by location in both sign and magnitude. Accounting for firm-level heterogeneity in production processes and policy responses, as well as the use of empirically grounded counterfactual baselines, can improve the realism of ex ante analysis and thus provide a more reliable basis for policy design.