I4R: The Power of Policy Evaluation

Statistical power and accumulating bias in evidence-based policymaking

This project examines how reliable the evidence behind policy decisions really is, using carbon pricing as a case study. The evidence-based policy paradigm typically relies on the idea that the accumulation of evidence from causal impact evaluations informs policy, for example via meta-analysis. But does combining evidence from multiple studies necessarily strengthen our knowledge, or can it reinforce existing biases?

Why does this matter?

Policy interventions are often evaluated using quasi-experimental designs that are vulnerable to bias. Such biases do not necessarily cancel out when results are aggregated in meta-analyses. Limited statistical power compounds this problem: policy effects are often small relative to background variation, making them difficult to detect reliably, even with large datasets.

What we're doing

The project uses the influential meta-analysis of carbon pricing by Döbbeling-Hildebrandt et al. (2024) as an entry point to the evidence on this policy measure. We, first, critically reassess the author’s treatment of statistical power. Second, we scrutinize the evidence base with respect to reproducibility and robustness.

What comes next

We systematically replicate influential studies included in Döbbeling-Hildebrandt et al. (2024) to assess the reproducibility and robustness of the underlying evidence and explore the implications for evidence synthesis more broadly.

Project outputs

Piseddu et al. (2026). The power of carbon pricing – A comment on Döbbeling-Hildebrandt et al. (2024) and its press release. I4R Discussion Paper No. 283. https://www.econstor.eu/handle/10419/338627

Fernando, Sachintha (2026): Are the Effects of Finland’s Carbon Tax on Transport Emissions Robust to Donor Pool Composition?, I4R Discussion Paper Series, No. 317. https://www.econstor.eu/bitstream/10419/342986/1/I4R-DP317.pdf