Services vs. Subsidies: Quantifying hassle costs in energy efficiency retrofits

The intervention is an online survey experiment conducted among household heads living in owner-occupied homes in Germany. The experiment is implemented as part of the fifth wave of the German Heating and Housing Panel (GHHP). Eligibility for the main experiment is restricted to respondents living in older homes built who have not replaced their heating system since construction. Respondents who have recently commissioned a heating system replacement are assigned to a separate supplementary module.

2025-10-13

2025-11-02

Our primary outcomes are as follows (in no particular order):

- Incentivized valuation of a hassle-reduction service (subsidy-equivalent value) for adopting an energy efficient heating system
- Beliefs about hours saved by the service
- Attribution of service value: time-savings vs. residual value
- Implied hassle cost per hour

We construct our primary outcomes as follows:

- Incentivized valuation (SEV): This is derived from Round-2 multiple price list (MPL) choices between a subsidy and the randomized service. We left-censor the SEV at €50 if the respondent always chooses the subsidy and right-censor using the mean reported follow-up amount if the respondent always chooses the service up to €5,100. For all other respondents, we identify the switching point and map interval responses to their midpoints. Inconsistent response patterns are set to missing.

- Beliefs about hours saved by the service: We convert the categorical responses from the eight-point scale into numeric values using midpoints (e.g., “1–3 hours” → 2 hours). One day is defined as eight hours, following the survey instructions. We censor the resulting variable at 0 hours (left) and 40 hours (right).

- Attribution of service value: time-savings vs. residual value: We compute the stated share of each respondent’s SEV that can be attributed to time-savings in two steps. First, respondents rate the relative importance of time-savings versus risk reduction on a five-point scale, which we translate to 100%, 75%, 50%, 25%, and 0% of SEV attributable to time-savings (options 1–5/6 respectively). Second, respondents indicate whether other (“residual”) factors influenced their valuation and rate their importance on a separate five-point scale. Combining both allows us to estimate the proportion of total SEV that each respondent attributes to time-savings.

- Implied hassle cost per hour: We calculate the implied hourly hassle cost by dividing the portion of SEV attributed to time-savings by the respondent’s believed number of hours saved by the service.

Our secondary outcomes are as follows (in no particular order):

- Incentivized willingness to accept a subsidy for adopting an energy efficient heating system
- Beliefs about total hours of adoption hassle
- Total monetized hassle cost
- Beliefs about the risk of major problems during adoption
- Beliefs about costs, potential savings, and effects on property value of adoption
- Socioeconomic and building covariates

We construct our secondary outcomes as follows:

- Incentivized willingness to accept a subsidy for an energy efficient heating system: This measure is derived from the Round-1 multiple price list (MPL) choices between a subsidy and a fixed €50 cash payment. Based on the switching point, we classify respondents into three types: always-takers (choose the subsidy even at €100 or below), marginal types (switch between €600 and €5,100), and never-takers (always prefer the €50 cash). The offered subsidy is conditional on receiving the government subsidy for energy efficient heating system, administered by the KfW.

- Beliefs about total hours of hassle for adoption: Respondents report expected hassle on a seven-point scale. We convert categories to midpoints and translate days and weeks into hours following the survey instructions. This question is asked separately for respondents who have not yet conducted the replacement and for those who have.

- Total monetized hassle cost: We compute this by multiplying the estimated hourly hassle cost by the total believed number of hassle hours. If hourly hassle costs vary across tasks, we use their weighted average based on task relevance.

- Beliefs about the risk of major problems during the adoption: We elicit perceived probabilities of major problems on a five-point scale. For respondents who have already adopted an energy efficient heating system, we additionally record whether such problems occurred and which types (e.g., cost overruns, damages, delays).

- Beliefs about costs, potential savings, and effects on property value of adoption: We convert reported intervals for costs and expected monthly savings into numeric midpoints. The perceived effect on property value is coded as an ordered variable.

- Socioeconomic and building covariates: We collect detailed demographic and building information, including income, age, gender, education, household composition, building age, size, energy system, and indicators of retrofit intentions or steps already taken.

We randomize respondents in the main experiment into one of three potential hassle-reduction services (planning/bureaucracy/supervision), with equal probability (1/3 each). For respondents in the supplementary module no randomization or incentivization is conducted.

Respondents in the main experiment are randomly assigned with equal probability (1/3) to one of three hassle-reduction service treatments:

1. Planning service, which provides professional planning and coordination support;
2. Bureaucracy (subsidy-filing) service, which handles administrative, subsidy-related paperwork; and
3. Supervision (construction-monitoring) service, which provides quality control and oversight during implementation.

Each respondent thus faces one of three possible services. The assigned service is only described before the second valuation task to ensure comprehension. The randomization is performed automatically within the survey software and is not visible to respondents.

Respondents complete two incentivized multiple price lists (MPLs): one eliciting their willingness to accept a subsidy for adopting an energy efficient heating system, and one eliciting their valuation of the assigned hassle-reduction service relative to the same subsidy, both pertaining to adopting an energy efficient heating system (e.g. a heat pump). Both MPLs are implemented using the random real-payment mechanism described in the survey, ensuring incentive compatibility.

Computerized random number generation by forsa online platform.

Individual household heads of owner-occupied homes (respondents).

No

Between 3,750 to 4,500 clusters.

We expect approximately 7,500 respondents to begin the experiment. We expect around 50% to 60% to meet the main eligibility criterion (i.e., living in a home with an old heating system), resulting in an effective sample size of about 3,750 to 4,500 participants for the main experiment. Roughly another 30% to 40% are expected to qualify for the supplementary module (i.e., those that already installed a new heating system), yielding an additional sample of roughly 2,250 to 3,000 respondents for the supplementary module.

Approx. 1,300 to 1,500 respondents for each type of service (i.e., Planning, Bureaucracy, Supervision).

See PAP.