Bitcoin Mining Energy Demand: Price Elasticity and Cross-Country Panel Analysis
Causal inference pipeline (DoWhy, Pearl's do-calculus) recovers a backdoor-adjusted price elasticity of +0.119 for bitcoin mining energy demand; 106-country monthly panel confirms cross-sectional heterogeneity.
About This Project
This work was completed at the Centre for International Governance Innovation (CIGI). A policy brief based on the findings is currently under peer review. Once published, it will be linked here.
The full analysis and write-up are not available on this page ahead of publication. The results below reflect what I can share in the interim.
Findings
An end-to-end causal inference pipeline using DoWhy and Pearl’s do-calculus framework estimated the price elasticity of bitcoin mining energy demand. Key results:
- Backdoor-adjusted price elasticity: +0.119 (global weekly series, n = 314; all three refutation tests passed)
- 106-country monthly panel (n = 6,021) constructed from CBECI/CCAF data; estimated via panel OLS with entity fixed effects and clustered standard errors; direct price effect −0.368 (p < 0.001, R² = 0.995)
- Mediation decomposition through the hashrate channel; asymmetric price shock response test: Wald F(1,236) = 44.57, p < 0.001
The refutation tests (random common cause, data subset, placebo treatment) all returned estimates consistent with a genuine effect rather than a specification artifact.
Technical Notes
- Data: Cambridge Bitcoin Electricity Consumption Index (CBECI), Cambridge Centre for Alternative Finance (CCAF)
- Framework: DoWhy, Pearl’s structural causal model, DAG-encoded confounders