Journalpublished

The Rebound Effect and Energy Efficiency Policy

Kenneth Gillingham (Yale University), David Rapson (UC Davis), Gernot Wagner (Harvard University)
T
Curated by Ted Lango
Published May 9, 2026Updated May 10, 2026
Read Original ↗

Abstract

This paper clarifies conceptual issues around the rebound effect and reviews the empirical evidence. A key distinction is made between rebound from a costless exogenous improvement in energy efficiency versus rebound from a specific policy intervention, which have different welfare implications and empirical magnitudes.

Curator Summary

Gillingham, Rapson & Wagner make a distinction that's critical for WFM practitioners: the rebound from a costless efficiency improvement (AI handles contacts faster) is different from rebound from a policy intervention (we deploy AI to reduce headcount). The former is about consumer behavior response; the latter includes organizational and strategic responses. Most AI business cases conflate the two. The paper also clarifies that 'backfire' (>100% rebound) is rare at the micro level but the economy-wide picture is murkier.

Why It Matters

When modeling AI's impact on contact volume, you need to distinguish between the direct rebound (customers contact more because it's easier) and the strategic rebound (the organization deploys AI specifically to cut costs, triggering additional behavioral and structural responses). This distinction affects whether you use the 10-30% direct rebound range or the potentially larger total rebound that includes organizational responses.

Caveats

Energy economics context — the policy distinction needs careful translation to service operations. The paper is more conceptual than empirical for the specific question of service demand rebound. Published in 2016, before the current wave of generative AI deployment.

Discussion

No comments yet. Start the discussion.

Loading...