The open problem originates in Rau, Moran, Manton and Goggins (2020), who monitored 20 Irish social-housing households before and after whole-house retrofit but could not causally attribute the measured energy changes (no contemporaneous control group), could not separate reduced-heat-loss savings from behavioural substitution with whole-house metering alone, and observed only one post-retrofit heating season. The open question is to quantify, with credible causal identification and end-use-level measurement, the realized net energy savings of residential retrofit in occupied dwellings — decomposed into engineering savings, direct rebound (comfort-taking and fuel or end-use substitution), and indirect rebound (re-spending of bill savings) — and to establish whether these persist, intensify, or revert across multiple heating seasons. The relevant literature supplies these components separately — a large U.S. randomized evaluation (Fowlie, Greenstone and Wolfram 2018), staggered difference-in-differences persistence estimates for England and Wales (Peñasco and Díaz Anadón 2023), rebound reviews, and performance-gap reviews — but no study combines causal identification, end-use decomposition, both rebound channels, and multi-season persistence.

Background

Residential energy retrofit — upgrading an existing dwelling's building fabric (wall, roof and floor insulation, airtightening, window replacement) and sometimes its heating system to reduce energy demand — is a pillar of European and North American climate policy. Engineering calculations, such as Ireland's Dwelling Energy Assessment Procedure (DEAP, the national standard methodology underlying a dwelling's energy rating), routinely predict substantial savings from these measures, yet metered consumption in occupied homes often falls short of predictions. This shortfall is known as the energy performance gap (Cozza et al., 2021), and it is compounded by the prebound effect — the observation that occupants of thermally poor dwellings often consume less energy before retrofit than calculations assume, because they under-heat their homes (Sunikka-Blank and Galvin, 2012). Two behavioural mechanisms can erode savings after retrofit. Direct rebound arises when the effective price of warmth falls and occupants take more of the energy service — heating more rooms, longer, or to higher temperatures — or substitute one fuel or end-use for another (for example, heating water with an electric immersion heater instead of the retrofitted boiler system). Indirect rebound arises when the money saved on energy bills is re-spent on other goods and services whose production consumes energy (Druckman et al., 2011); Rau et al. (2020, Section 5.1) themselves record that it is "entirely unclear" whether post-retrofit bill savings are spent on more or less energy-intensive activities.

The behavioural and social side of the problem was framed by the Energy Cultures Framework (ECF) of Stephenson et al. (2010), which analyses household energy demand as the interplay of material conditions (the dwelling and its equipment), norms (expectations about comfort and adequacy), and practices (what occupants actually do). Rau, Moran, Manton and Goggins (2020) applied this lens to 20 households in an Irish social-housing estate — rental housing owned and managed by a local authority or housing association — before and after a whole-house energy retrofit. For each household they combined monthly whole-house gas and electricity meter readings with multi-room temperature and humidity logging and with attitude-and-practice surveys, across one pre-retrofit six-month window (February–July 2015) and one post-retrofit window (February–July 2016). They observed large house-to-house variation and indications of comfort-taking, but their design left the central quantities unidentified: with no contemporaneous control group, measured changes cannot be causally attributed to the retrofit rather than to weather, prices, occupancy changes, or other temporal influences; with whole-house metering only, savings from reduced heat loss cannot be separated from changes in how energy is used; and with a single post-retrofit heating season, persistence is unknown.

A neighbouring economics literature has shown that credible causal attribution is feasible. Fowlie, Greenstone and Wolfram (2018) used a randomized encouragement design — households were randomly assigned to receive outreach urging them to apply for the programme — into the U.S. Weatherization Assistance Program (a federal programme funding efficiency upgrades for low-income households) across more than 30,000 households, finding realized savings roughly 2.5 times smaller than engineering projections and no detectable increase in indoor temperature in a one-time follow-up winter survey. Peñasco and Díaz Anadón (2023) used a staggered difference-in-differences design — comparing gas-consumption trajectories of dwellings that installed insulation at different times — over five years before and after loft and cavity-wall insulation in England and Wales, finding savings that persist but whose magnitude remains contested. What no study yet provides is the combination that the retrofit-evaluation literature keeps calling for (e.g. Medrano-Gómez et al., 2025): causal identification, end-use-level measurement, quantified direct and indirect rebound, and multi-season persistence within one design, especially in tenanted social housing where occupants' scope for behavioural change is institutionally constrained.

Problem Statement

The design of Rau et al. (2020, DOI 10.1016/j.scs.2019.101983) left the central quantities unidentified, and the paper records the needed advances as distinct stated limitations and future-research directions (Sections 4.1.5, 4.2, 5.1 and 5.2). The open empirical question is: determine — with credible causal identification — how much of the engineering-predicted energy saving from residential retrofit is actually realized in occupied dwellings, and what happens to it over time. The question spans three dimensions, each grounded in a limitation the source itself states. (1) Causal attribution: separate the retrofit's effect on household energy use from weather, energy prices, occupancy changes and other temporal influences. (2) Decomposition: resolve the net change in energy use into savings from reduced heat loss (the engineering effect), direct behavioural rebound (increased comfort-taking and fuel or end-use substitution, such as replacing boiler heating with electric immersion heating), and indirect rebound (re-spending of bill savings on other energy-intensive goods and services). The in-dwelling components — the engineering effect and direct rebound — are to be separated from one another rather than treated as a single net change; indirect rebound is to be quantified from the re-spending of bill savings. (3) Persistence: establish whether the net savings and the behavioural adjustments persist, intensify, or revert across multiple heating seasons. The setting of greatest interest is tenanted social housing, where institutional arrangements (landlords, housing associations, tenancy terms) mediate occupants' scope for behaviour change, but the question is posed for occupied dwellings generally, with quantified uncertainty throughout.

References

  1. Rau, H., Moran, P., Manton, R., & Goggins, J. (2020). Changing energy cultures? Household energy use before and after a building energy efficiency retrofit. Sustainable Cities and Society, 54, 101983. DOI: 10.1016/j.scs.2019.101983, https://doi.org/10.1016/j.scs.2019.101983
  2. Stephenson, J., Barton, B., Carrington, G., Gnoth, D., Lawson, R., & Thorsnes, P. (2010). Energy cultures: A framework for understanding energy behaviours. Energy Policy, 38(10), 6120–6129. DOI: 10.1016/j.enpol.2010.05.069, https://doi.org/10.1016/j.enpol.2010.05.069
  3. Fowlie, M., Greenstone, M., & Wolfram, C. (2018). Do energy efficiency investments deliver? Evidence from the Weatherization Assistance Program. The Quarterly Journal of Economics, 133(3), 1597–1644. DOI: 10.1093/qje/qjy005, https://doi.org/10.1093/qje/qjy005
  4. Peñasco, C., & Díaz Anadón, L. (2023). Assessing the effectiveness of energy efficiency measures in the residential sector gas consumption through dynamic treatment effects: Evidence from England and Wales. Energy Economics, 117, 106435. DOI: 10.1016/j.eneco.2022.106435, https://doi.org/10.1016/j.eneco.2022.106435
  5. Peñasco, C., & Díaz Anadón, L. (2024). A comment on "Assessing the effectiveness of energy efficiency measures in the residential sector gas consumption through dynamic treatment effects: Evidence from England and Wales". Energy Economics, 139, 107946. DOI: 10.1016/j.eneco.2024.107946, https://doi.org/10.1016/j.eneco.2024.107946
  6. Cozza, S., Chambers, J., Brambilla, A., & Patel, M. K. (2021). In search of optimal consumption: A review of causes and solutions to the Energy Performance Gap in residential buildings. Energy and Buildings, 249, 111253. DOI: 10.1016/j.enbuild.2021.111253, https://doi.org/10.1016/j.enbuild.2021.111253
  7. Sunikka-Blank, M., & Galvin, R. (2012). Introducing the prebound effect: the gap between performance and actual energy consumption. Building Research & Information, 40(3), 260–273. DOI: 10.1080/09613218.2012.690952, https://doi.org/10.1080/09613218.2012.690952
  8. Cholewa, T., Balaras, C. A., Nižetić, S., & Siuta-Olcha, A. (2020). On calculated and actual energy savings from thermal building renovations – Long term field evaluation of multifamily buildings. Energy and Buildings, 223, 110145. DOI: 10.1016/j.enbuild.2020.110145, https://doi.org/10.1016/j.enbuild.2020.110145
  9. Mashhadi Rajabi, M. (2022). Dilemmas of energy efficiency: A systematic review of the rebound effect and attempts to curb energy consumption. Energy Research & Social Science, 89, 102661. DOI: 10.1016/j.erss.2022.102661, https://doi.org/10.1016/j.erss.2022.102661
  10. Druckman, A., Chitnis, M., Sorrell, S., & Jackson, T. (2011). Missing carbon reductions? Exploring rebound and backfire effects in UK households. Energy Policy, 39(6), 3572–3581. DOI: 10.1016/j.enpol.2011.03.058, https://doi.org/10.1016/j.enpol.2011.03.058
  11. Dineen, D., & Ó Gallachóir, B. P. (2017). Exploring the range of energy savings likely from energy efficiency retrofit measures in Ireland's residential sector. Energy, 121, 126–134. DOI: 10.1016/j.energy.2016.12.024, https://doi.org/10.1016/j.energy.2016.12.024
  12. Mac Uidhir, T., Rogan, F., Collins, M., Curtis, J., & Ó Gallachóir, B. P. (2020). Improving energy savings from a residential retrofit policy: A new model to inform better retrofit decisions. Energy and Buildings, 209, 109656. DOI: 10.1016/j.enbuild.2019.109656, https://doi.org/10.1016/j.enbuild.2019.109656
  13. Morgan, D. J., Maddock, C. A., & Musselwhite, C. B. A. (2024). These are tenants not guinea pigs: Barriers and facilitators of retrofit in Wales, United Kingdom. Energy Research & Social Science, 111, 103462. DOI: 10.1016/j.erss.2024.103462, https://doi.org/10.1016/j.erss.2024.103462
  14. Taptiklis, P., Fyfe, C., Grimes, A., Minehan, S., & Benison, T. (2026). A New Zealand heat pump retrofit programme: Warmer, Drier, Happier, Cheaper. Journal of the Royal Society of New Zealand, 56(2), e70043. DOI: 10.1002/snz2.70043, https://doi.org/10.1002/snz2.70043
  15. Medrano-Gómez, L. E., Boarin, P., & Premier, A. (2025). The retrofit puzzle: Connecting practices, retrofit measures, and performance outcomes through socio-technical evaluations. Energy Research & Social Science, 120, 103924. DOI: 10.1016/j.erss.2025.103924, https://doi.org/10.1016/j.erss.2025.103924
  16. Curtis, J., & Farrell, N. (2025). Irish residential heating demand profile dataset. Data in Brief, 58, 111215. DOI: 10.1016/j.dib.2024.111215, https://doi.org/10.1016/j.dib.2024.111215
  17. Huaccha, G. (2023). Regional persistence of the energy efficiency gap: Evidence from England and Wales. Energy Economics, 127, 107042. DOI: 10.1016/j.eneco.2023.107042, https://doi.org/10.1016/j.eneco.2023.107042