Steering Homeowners Toward Net Zero: Experimental Evidence on Retrofit Loan Design and Benefit Communication

Participants are shown a loan quote for a home energy upgrade under one of six experimental conditions, varying loan duration and the type of benefit information provided alongside the quote.
The six conditions are:
1. T1: 5-year loan, energy savings information only
2. T2: 10-year loan, energy savings information only
3. T3a: 5-year loan, energy savings plus information on potential property value increases
4. T3b: 5-year loan, energy savings plus information on health benefits
5. T4a: 10-year loan, energy savings plus information on potential property value increases
6. T4b: 10-year loan, energy savings plus information on health benefits
The intervention is delivered in two sequential scenarios. In the first scenario, all participants, regardless of their assigned condition, are shown a standardised loan quote for the installation of a heat pump in a semi-detached house, accompanied by energy savings information only, under their assigned loan term (5-year or 10-year). This scenario holds retrofit type, dwelling type, and information type constant across all participants, with the only variation being loan duration.
All participants are shown their estimated annual energy savings, expressed as a percentage, based on their retrofit type and dwelling type. The interest rate is held constant across all conditions at 5% APR. Loan amounts, repayment schedules, and benefit estimates reflect currently available retrofit schemes and costs in Ireland.
In the second scenario, participants use an interactive calculator to generate a personalised loan quote by selecting their own property type and preferred retrofit option(s) — attic insulation, double-glazed windows, solar PV, a heat pump, or external wall insulation — and indicating whether they wish to apply a government/SEAI grant. A personalised loan quote is then generated reflecting their selections, their assigned loan term, and their assigned information condition, as specified by the six conditions above.

2026-05-21

2026-07-31

Primary Outcomes–
1. Intention to retrofit: A binary variable (1/0) indicating whether the participant reports an intention to undertake a retrofit, given the loan terms and information presented.
2. Intention to take up the retrofit loan: A binary variable (1/0) indicating whether the participant reports willingness to avail of the retrofit loan under the terms presented.
3. Engagement with further retrofit information: A binary variable (1/0) indicating whether the participant opts to read additional information about retrofit loans, with the survey noting this would add approximately 2 minutes to their completion time.

1. Willingness to recommend retrofitting: Whether the participant would recommend retrofitting to a friend or family member, given the loan options and information presented.
2. Perceived affordability: The participant's assessment of how affordable the retrofit loan feels under the presented terms.
3. Perceived overall benefits: The participant's assessment of the overall benefits of retrofitting, encompassing energy savings, potential property value increases, and health improvements.

Research Questions and Hypotheses
This study addresses two research questions. First, does offering a longer loan term (10 years vs. 5 years) increase homeowners' intention to retrofit and their willingness to take up a retrofit loan? Second, does providing co-benefit information on potential property value increases or health improvements, alongside energy savings information, increase retrofit intention and loan uptake relative to presenting energy savings information alone?
We test the following hypotheses:
H1: Participants assigned to a 10-year loan term will report higher retrofit intention and higher willingness to take up the loan than those assigned to a 5-year term, driven by lower perceived monthly repayment burden (T2 vs T1).
H2: Participants shown co-benefit information (property value or health benefits) alongside energy savings information will report higher retrofit intention and higher willingness to take up the loan than those shown energy savings information only, driven by higher perceived overall gains from retrofitting (T4 vs T3).
H3: The effects of property value information and health benefit information on retrofit intention and loan uptake will differ from one another, reflecting differences in the salience and perceived relevance of each co-benefit type (T4b vs T4a and T3b vs T3a).
Sample
Participants are recruited through a reputable Irish survey firm. The sample is restricted to homeowners and is designed to be nationally representative with respect to age, region, and gender. The total planned sample size is 2,900 participants, allocated approximately equally across the four primary treatment arms (~725 per arm), with the co-benefit arms (T3 and T4) each split evenly into two subgroups of approximately 362–363 participants.
Experimental Flow
Participants are first randomised into two groups based on loan duration: a 5-year loan group and a 10-year loan group. All participants in both groups are then shown a standardised loan quote for the installation of a heat pump in a semi-detached house, based on their assigned loan term. Participants indicate whether they would intend to retrofit and whether they would take out the loan under these fixed conditions. This standardised scenario ensures that retrofit type and property type are held constant across participants, allowing clean identification of the effect of loan duration on retrofit decisions independent of individual preferences.
Following the standardised scenario, all participants proceed to a tailored scenario in which they use an interactive calculator to generate a personalised loan quote based on their own property type and preferred retrofit option(s), under their assigned loan term and information condition. In the tailored scenario, within each loan-duration group, participants are further randomised into two subgroups based on the type of benefits information: energy savings information only or energy savings plus additional co-benefit information. Together, the two scenarios allow treatment effects to be identified under controlled, standardised conditions and under more realistic, preference-driven conditions.
Participants assigned to the co-benefit arms (T3 and T4) are further randomised into two subgroups, one receiving additional information on potential property value increases only, and the other receiving additional information on health benefits only.
The expected causal mechanisms are twofold. First, longer loan terms lower monthly repayments, increasing perceived affordability and thereby raising both intention to retrofit and willingness to take up the loan. However, it could also be that retrofit and loan take-up are higher with shorter loan terms since longer-term loans are associated with a higher total cost of credit, and participants put more weight on the total cost compared to the monthly repayment amount. Second, exposure to property value or health co-benefit information increases perceived overall gains from retrofitting, strengthening motivation and leading to higher stated retrofit intention and loan uptake.

Not available

Randomisation done by a computer program

Individual

No

No clusters

2900 individuals

725 per primary treatment arm; groups (3) and (4) are each split into two subgroups of approximately 362–363 participants to allow separate identification of property value and health co-benefit effects.

The analysis that compares treatment groups T2 and T4 (n=725 each) against their respective control groups, T1 and T3 (n=725 each), for which the study is expected to achieve a minimum detectable effect size of 7.14 percentage points with 80% power at a Bonferroni-corrected significance level of α=0.025 (adjusting for two primary hypotheses, T1 vs T2 and T3 vs T4). The analyses that examine the differential effect of property value vs health benefits as co-benefit information (n=362 vs n=362) will have a minimum detectable effect size of 9.64 percentage points with 80% power at α=0.05.