1.1 Setting out the pathway for decarbonising heating systems in existing buildings

Given the climate crisis and the delay in achieving climate policy objectives, there is an urgent need to shift away from oil and gas-based heating systems in buildings (representing 26% of global energy-related CO₂ emissions and 34% in the EU; [1,2]). Old buildings are characterised by much higher energy requirements for space heating than new ones and by a dominant share of pre-existing fossil fuel-based heating systems (oil and gas boilers) which, once having reached the end of their lives, are typically again replaced by oil and gas boilers [3]. The European Union’s (EU) legislation foresees the entire building stock to be fully decarbonised by 2050. While all new buildings in the EU must be zero emission buildings as of 2030 (and they are required to be “nearly zero-energy buildings” already today), the focus of the EU’s Building Directive [4] is primarily on energy performance (i.e., reduction of specific energy use in kWh/m²/a), thereby prioritising action on the least efficient buildings. As complementary approach, the EU’s Energy Efficiency [5]) defines for the period 2021–2030 mandatory final energy savings which include but are not limited to buildings (Article 8); it also calls for the preparation of local heating and cooling plans (for municipalities with more than 45000 inhabitants) and it specifies decarbonisation levels for thermal grids for the period 2025–2050, while not specifically addressing decentralised heat supply. The EPBD requests the implementation of national policies and measures for improving the energy performance of buildings and it calls for “a clear legal basis” for banning fossil fuel boilers, however without providing guidance (e.g., on the pathway or by defining any requirements). Neither do the Regulation on EU Governance and Climate Action [6] and nor the European Climate Law [7] provide any specific guidance on the decarbonisation of buildings. Against this background, decarbonisation in particular of heating systems in existing buildings in countries with temperate climate is still a “formidable policy challenge” [8].

1.2 Renewable energy heating options

Next to heat pumps which represent the focus of this paper, there are a number of other renewable heating solutions, namely bioenergy, solar thermal energy and geothermal energy. For pollution reasons (esp. particulate matter), bioenergy in the form of wood logs, chips or pellets is restricted or even prohibited for building heating systems in some countries and urban areas [9] while wood and pellets are increasingly used for district heating systems (e.g., [10]), typically in combination with pollution prevention measures (centralised systems can be equipped with appropriate filters). Rural areas often promote the use of locally available wood due to the lower cost, to support the local economy and to increase self-sufficiency while reducing greenhouse gas emissions. Biogas represents another suitable bioenergy but given its wide applicability including transport and high-temperature processes in industry next to buildings, the potential use of biogas exceeds by far its maximum availability (cf. [11]). Green hydrogen, i.e., hydrogen (H₂) generated using renewable electricity, is sometimes also proposed as option for heating buildings, thereby using the existing natural gas infrastructure [12,13], while the counterarguments put forward against this technology are the low conversion efficiency from electricity to H₂ and the resulting need for disproportionately large amounts of renewable electricity as well as the high cost. Green hydrogen is consequently not a large-scale, short-term solution for heating buildings. Waste heat, e.g., from municipal waste incineration plants or industrial processes, is a further option but also here the availability is a limiting factor, next to the need of a district heating infrastructure. In dense urban areas, district heating is a preferred choice, with large-scale heat pumps representing an emerging heat source and complementing the use of waste heat and biomass as decarbonisation strategies. While more attention will need to be paid in future to district heating combined with large-scale, centralised heat pumps, decentralised heat pumps substituting oil and gas boilers are a short to medium-term solution for buildings without connection to district heating and represent the focus of this paper. Their implementation should be combined with prior improvement of the thermal performance of buildings in order to avoid high electricity demand in the cold season [14,15] which entails the risk of jeopardising electricity supply security. Given the climate urgency and the long time periods as well as the high investment costs associated with thermal retrofit of the building envelope, (decentralised) heat pumps replacing oil and gas boilers are increasingly being put forward as solution per se, in spite of the priority which should actually be given to first reduce the energy demand [16].

A heat pump that runs on 100% renewable electricity can be considered as convincing solution for decarbonisation. This requires that renewable electricity is available at any moment at which the heat pump is operated, i.e., also in winter, which is likely to require some seasonal energy storage (e.g., using power-to-gas technologies; [17,18].

1.3 Mixed policy experience to date

Since the advantages of heat pumps and of some other renewable heating technologies are too limited to trigger their autonomous, large-scale diffusion [8], dedicated policies are required. [19] prepared an overview of phase-out regulations for fossil fuel heating in EU Member States, showing that only few countries have started to implement more coercive measures: In Denmark, there is an obligation for renewable heating as default but with exceptions depending on the zone (e.g., heating with natural gas is still possible in areas served by a gas grid). Other countries focus on the phase-out of fuel oil (Belgium, France, Greece, Slovenia), ban the use of fossil fuels in new buildings (Austria, Belgium, Ireland, Slovenia) and/or foresee distant deadlines for decommissioning fossil fuel boilers (e.g., 2035–2040 in Austria). In 2024, Germany implemented a 65% renewable energy share for newly installed heating systems in existing buildings as of mid 2026 and mid 2028 (depending on the size of the municipality). However, the new government in Germany, has announced to revise this law, thereby offering more technology options. Similarly, in the Netherlands, the former government had issued a de facto ban for pure natural gas boilers as of 2026 which was, however, cancelled by the new government [20,21]. Full decarbonisation of heating in existing buildings is hence not prescribed in any EU country in the short term while the most demanding decarbonisation targets were foreseen in Germany but remained by far less ambitious than in Switzerland and the German requirements will probably be further reduced. Overall, the gap between actual decarbonization of buildings in the EU and the target pathway has continually increased, making it increasingly challenging to reach climate neutrality in the building sector by 2050 [22].

1.4 Ensuring policy effectiveness

In order to create the intended impact, policies need to be accepted in the first place. As the experience outlined above shows, this is not a given. Second, when being implemented, they need to be effective. The true effectiveness of energy and climate policy measures is often critically discussed with reference to the rebound effect (e.g., [23,24]) or as a consequence of other reasons for increased activity such as economic growth [25]. For heat pumps, there are so far only very few papers on the rebound effect [26]. Most research on the rebound effect typically quantifies its size for empirical cases in the form of a demand elasticity [27]. In absence of the data required to conduct this type of analysis for heat pumps, it is worthwhile to categorise different subsidy levels and different beneficiaries and to reflect about the related implications.

1.5 Switzerland as experimental ground to learn from

Similar to the EU, new buildings in Switzerland should be exclusively supplied with renewable heat and at latest by 2050, no building should emit CO₂ (EnDK Principle 2; Grundsatz 2, Principe 2; [28]). To this end, the necessity of defining maximum levels of non-renewable energy used for heating was stipulated in Swiss national law (Federal Energy Law, Eng (SR 730.0), Art. 45). The concrete maximum levels are specified by cantonal legislation because in Switzerland, buildings including their energy use are subject to cantonal laws which is a feature of the prevailing federalist system (see S2 Text for details). While energy and climate policies are harmonised to some extent across the cantons by means of so-called cantonal model regulations (CMR; known as MuKEn in German and MoPEC in French), there is leeway for significant differences, offering possibilities for comparative analyses which cannot be conducted in most other countries. The last two CMRs were released in 2014 (therefore referred to as CMR2014) and in 2025 (CMR2025). Since CMR 2025 was published after the completion of the research presented in the present paper, we discuss in the following CMR2014 while we will revert to CMR 2025 in section 4 (discussion). CMR2014 foresaw that at least 10% of the energy used by new heating system must be renewable. While originally implemented for incremental implementation of renewable energy (e.g., solar thermal systems for domestic hot water), some cantons have advanced more quickly: they implemented bans or quasi-bans for the replacement of oil and gas boilers in existing buildings, which typically represent the preferred choice if no dedicated policy is in place (e.g., prior to the new policy, in 84% of all building owners in the city of Zurich again choose an oil or gas boiler when replacing their heating system; [3]).

In Switzerland, the installation of heat pumps is a rather convincing solution in terms of greenhouse gas abatement because local electricity is characterised by a very low carbon footprint due to its primary reliance on hydropower and nuclear energy, even when accounting for electricity imported primarily from Germany and especially in winter [29].

1.6 Objectives of this study

Against this background the objective of the present paper is to compile and critically discuss the experience made with policy measures for the decarbonisation of building heating systems in Swiss cantons which are leaders in this domain. Given their importance for the Swiss energy transition, the focus is on heat pumps for the replacement of oil and gas boilers. Even if, in Switzerland, the transition to heat pumps is less advanced for existing multi-family buildings than for existing single-family houses, the early experience made can nevertheless inform policymakers and other stakeholders in other countries characterised by temperate climate and aspiring (or already having) a low carbon footprint for electricity. Current insights on costs of heat pump systems in Switzerland also allow to better understand the financial viability which strongly influences acceptance of energy policy measures [30] and raises questions about their distributional impacts [31].

Next to implementation challenges, potential rebound effects associated with the design of the policy measure are discussed. Section 2 briefly describes the approach and methodology and section 3 presents results which are critically discussed in section 4, prior to the conclusions in section 5.

3.1 Main policies implemented

According to the interviews, the three main policy instruments that have been implemented for decarbonising building heating systems are i) information campaigns, ii) regulation mandating a minimum share of renewable energy for space and water heating when replacing an existing fossil fuel-based boiler, and iii) subsidies:

• Information and advice: At the national level, the publicly funded programme RenewableHeating offers publicly available information (technical information, on-line cost calculation tool as well as information on financing and subsidies, organisational steps and practical cases) as well as initial free-of-charge personal advice about renewable heating systems to building owners (private, public, enterprises; [ 21,39]). A major effort is also made by the cantonal offices to inform the various stakeholders about the energy and climate policy objectives, technological solutions (independent advice) and available subsidies. The main stakeholders addressed are homeowners, real estate companies, installers, energy experts consulting homeowners, architects and engineers. The types of information channels range from public evening events, to stands in public spaces, telephone information services (e.g., from 8 am to 7 pm in Lucerne) and co-financed energy audits (so-called CECB+ or GEAK+ providing recommendations for practical action both for the short term and the medium to longer term). One important purpose of information and advice is to create awareness about the need to switch to a renewable heating system when the existing boiler breaks down, to avoid unpleasant surprises, negligence and rejection. The local information campaigns (e.g., at the neighbourhood level) must be based on local heating plans, specifying future requirements and boundary conditions (see also below). To enable planning (organisational, financial) for owners but also the other stakeholders, rather long transition periods are foreseen (e.g., in Lucerne until 2040 depending on the specific case). Particular attention is paid to information for installers (e.g., Jura) and to training programmes for them (Lucerne) because building owners tend to follow their advice.
• Regulatory measures: In all cantons consulted, the regulatory measures focus on the replacement of old or deficient heating systems requiring replacement, while a dedicated policy for early retrofit is additionally supported by the city of Zurich (complementary to the support by the canton of Zurich). As shown in Table 1, the original 10% requirement for renewables as already defined by CMR2014 is currently only applied in one canton (until recently in three cantons) among the seven cantons studied. The canton of Thurgau increased the requirement only very slightly (from 10% to 15%), while the canton of Jura very significantly reframed its legislation. In Neuchatel, where there is also a minimum level of 20%, the full replacement by renewable energy (100%) is mandatory if this is technically and financially possible. In Zurich, the regulation is essentially the same, even though it is worded slightly differently (this is also the case for Basel-City). Interestingly, even when setting the share at only 10% or 20%, the cantons listed in Table 1 aim for a largely complete transition to renewable energy in existing buildings by implementing complementary policy measures, in particular by information campaigns and subsidies, often supported by simplification of the permit process as explained below. Nevertheless the cantonal policymakers expressed caution about defining too ambitious requirements for the renewable energy share and they were generally not in favour of banning fossil fuels since these could compromise policy acceptance and would require more preparation (energy planning, potentially implementation of district heating, communication etc.). At the same time, some of the cantonal policymakers indicated that the support for more stringent renewable energy requirements was growing, that the stringency of the polices should be gradually increased and that a full ban of fossil fuels may be possible in future (in the meanwhile we know that this expectation expressed by some of the policymakers in 2022 was realistic, as will be discussed in section 4).
• Subsidies: In six out of the seven cantons, subsidies are provided to (partly) compensate the higher investment costs of renewable energy technologies. Zurich is the only exception by establishing the subsidies based on the difference between the levelised cost of the renewable energy technology and the oil or gas based heating. As expected, the subsidies typically comply with the subsidy bracket defined by the so-called Harmonised Subsidy Model (HFM in German and ModEnHa in French; [41]).

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Minimum share of renewable energy sources required by cantonal energy law when replacing an existing heating system. Source: EnDK, implementation of CMR2014; cantonal announcements on subsidy levels (CMR2025 was announced in summer 2025 but cantons have not yet had the time to implement it).

https://doi.org/10.1371/journal.pclm.0000608.t001

In the following, we summarise the insights from the interviews with regard to the technical and financial viability of replacing conventional heating systems by heat pumps, the (perceived) acceptance of the policy and its estimated effectiveness, further enablers as well as views about bans of fossil fuel-fired boilers. While our main focus is on the replacement of fossil fuel-based heating systems we include also some aspects around the replacement of direct resistance heating the installation of which was promoted in some European countries in the 1960s, 1970s and partly also the 1980s. When interpreting the outcome of the interviews, it should be kept in mind that the experience made in the seven cantons is still relatively recent (see dates in Table 1) and that none of the cantons has officially confirmed, up-to-date data about the implementation level of renewable energy technologies including heat pumps and that they instead rely on estimates (building owners do not always report their data and municipalities do not always collect and report the data to the higher administrative level; one interviewee described the resulting quality of the associated data in the Swiss Building Register as good for one-third of the cases, medium for another third, and insufficient for the last third).

3.2 Viability of policies

3.3 System-wide CO₂ emissions

The efficiency of a heat pump system can be described by the annual Coefficient of Performance (COP_annual, also referred to as Seasonal Performance Factor) which represents the ratio of the annual amount of heat provided by the heat pump to its electricity use. Together with the CO₂-intensity of the national electricity supply, the Coefficient of Performance determines the CO₂ emissions of heat pumps. Fig 1 displays the CO₂ emissions of a heat pump as a function of COP_annual for the electricity mix in different European countries. As proxy for CO₂-intensity of the electricity used, the national electricity supply mix was considered (we will revert to this simplification). While, according to Fig 1, a COP_annual of 2.0 allows a decarbonisation by a factor of more than two thanks to the low CO₂ emission intensity of Swiss grid electricity, the same heat pump system operated in Germany (with today’s electricity mix) would lead to somewhat higher CO₂ emissions than a gas boiler which would not be acceptable. It is therefore necessary to reach COP_annual values of at least 2.5 or 3.0 while further decarbonising grid electricity. Overall, the figure shows that very low CO₂ emissions levels can be reached by heat pumps in case of near-zero electricity supply combined with high COP values.

It should, however, be noted that a number of simplifications were made in this analysis. First, instead of using the CO₂-intensity of the electricity supplied by the national power producers, the mix of consumed electricity accounting for imports and exports should have been considered (for example, Switzerland imports CO₂-intensive electricity from Germany in winter while it exports near-zero electricity in summer). While this type of analysis has been conducted for Switzerland [47,48,29] the analysis would need to be redone for the different COPs due to their different temporal heat demand patterns. In addition, no comparable analyses are available for other countries. As further simplifications, only the direct energy-related CO₂ emissions were considered while a comprehensive analysis would require a life cycle approach, thereby including the impacts of all supply chains including potential refrigerant losses. By addressing most of these points, the analyses by Rüdisüli et al. [29] show that significant savings can be achieved also according to a comprehensive analysis. In addition, it must be kept in mind that electricity trade deteriorates the carbon balance for some countries (e.g., Switzerland), while it reduces the CO₂ footprint for others (e.g., Germany).

3.4 Analysis of levelised cost

In this section, we aim to compare a typical situation for an electric air source heat pump with a gas boiler, thereby acknowledging that real datasets show wide ranges (see [33], for real data in Geneva and Basel city, thereby also studying ground source heat pumps). We assume here a residential building with a specific heat demand (useful energy) of 100 kWh per m² and a COP_annual value of 2.8. The data used in Table 2 (also available as downloadable Excel file with implemented calculations, see supplementary material) originates from the on-line cost calculation tool of RenewableHeating [ 21,39], with some exceptions, i.e., the subsidy was somewhat lowered (to better reflect the values reported in Table A in S2 Text) and the gas boiler investment cost was also somewhat reduced (from approximately 3000 CHF/kW_th to 2500 CHF/kW_th). As Table 2 shows, the investment cost for an air source heat pump is nearly 3-fold of the cost of the investment cost of a gas boiler, while even larger ratios have been reported in the past (M. Rüetschi, 2022). Investment costs for installed heat pump systems are similar regardless of the size because economies of scale of larger heat pumps are compensated by the more demanding technical integration (personal communication with M. Rüetschi, 2022). For the investment period of 20 years, the future electricity price is assumed to remain identical with the value of the previous three years (24 cents/kWh), whereas a markup of 13% was assumed for natural gas, resulting in a gas price of close to 16 cents/kWh. Data were collected in Swiss Francs but are reported in EURO since the values hardly differ (1 EUR_2021–2024 = 1.003 CHF_2021–2024 = 1.100 USD_2021–2024; values from European Central Bank, referring to the 4-year period from 1 Jan 2021 until 31 Dec. 2024).

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Levelised cost of an air source heat pump (without and with different levels of subsidy) and of a gas boiler.

https://doi.org/10.1371/journal.pclm.0000608.t002

Based on these assumptions the levelised cost (without subsidies) of the air source heat pump is 16% above the levelised cost of the gas boiler (Table, column I and V). For a typical subsidy for the heat pump (equivalent to 11% of the investment cost; see column II), its levelised cost remains in an acceptable range compared to the gas boiler (by 7% more expensive). In order to reach identical levelised cost, the subsidy would need to be raised to close to 20% (column III). Equivalent levelised cost can hence be achieved with reasonable subsidy levels. If, on the other hand, the level of the subsidy is chosen to ensure comparable investment costs, a 65% subsidy would be required and the levelised cost would be nearly 38% below the respective value of the Gas boiler for the chosen case (column IV and V). Comparison with Supporting information S2 Text on the subsidy levels of the seven studied cantons shows that only the very high support provided by Basel-City (9250 CHF for a 5 kW air source heat pump) would lead to similarly low net investment costs as in case IV.

The parameter analysis displayed in Fig 2 highlights the rather strong sensitivity to both the energy price ratio (the electricity price was kept at 24 cents/kWh while the gas price was varied) and to the investment costs (a typical subsidy of 800 EUR/kW_th was assumed for the heat pump while the gas boiler cost was varied). For example, as visible from Table 3 (which displays the data presented in Fig 2), for a typical subsidy level represented by a ratio of investment costs of 2.56 (HP/GB), only energy price ratios beyond 0.65 (Gas to electricity price ratio) ensure cost-effectiveness of the heat pump (i.e., a levelised cost ratio for the heat pump to the gas boiler below 1.0). Even for the rather optimistic case of an investment cost ratio (HP/GB) of 2.0, the energy price ratio (Gas/El.) would need to exceed 0.6. These values would require high investment subsidies and/or high CO₂ or fossil fuel taxes or otherwise strongly subsidized electricity prices for heat pump operation.

Download:

• PNG
  larger image
• TIFF
  original image

Table 3. Tabular representation of the data displayed in Fig 2.

https://doi.org/10.1371/journal.pclm.0000608.t003

3.5 Potential subsidy-related rebound effects

To assess rebound effects, we can distinguish different cases:

1. i. If the subsidy is based on identical levelised cost (for the heat pump and the fossil fuel boiler) and the building/apartment is used by the owner, then the total costs for owner remain unchanged (identical levelised cost) and therefore, no rebound effect is expected.
2. ii. If the subsidy is based on identical investment costs and the building/apartment is used by the owner, then this owner has more disposable income than in the previous case, which can be expected to result in a rebound effect. This also occurs for those owners who would anyway have installed a renewable heating system even without a subsidy (freeriding).
3. iii. If the property is rented to a tenant, with the latter paying the energy bill and in case the subsidy is based on identical investment costs, then no rebound should occur on the owner’s side but a rebound can be expected for the tenant due to lower energy bill.
4. iv. If, finally, the property is rented to a tenant, with the latter paying the energy bill and in case the subsidy for the owner is based on identical levelised cost then there is a rebound effect for the tenant (who has more income) but the owner has more expenses than before, resulting in a negative rebound. Whether the positive and negative rebound compensate each other depends on a variety of factors including the spending patterns.

It could be considered quantifying these rebound effects with economic input/output (I/O) analysis. We refrain from doing so in view of the uncertainties and difficulty to correct for other influencing factors and their effects. It can nevertheless be concluded that subsidies based on levelised cost help to avoid the rebound effect. If this is a primary objective, it would also be desirable to keep the energy costs for tenants unchanged. Reduced energy costs should, however, be prioritised in the case of socially disadvantaged tenants.

The abovementioned increased prices for installed heat pumps observed in the presence of subsidies (or their increase) can be considered as another type of rebound effect.

While the assessment of the four rebound cases considers avoided costs (which become available for additional expenses), it should also be considered what the origin of the funds is which are used for financing the subsidies (we may refer to this as prebound).

1. i. If the renewable energy subsidies are financed from tax revenue, this revenue could have been spent for other purposes (e.g., for education, health care or for tax cuts for employees and/or companies). In the theoretical case of a foregone tax cut for employees as a consequence of the subsidy the prebound can be expected to be small because of the similarity of the beneficiaries. Nevertheless distributional effects would be likely to occur.
2. ii. If renewable energy subsidies are financed by additional state debt (e.g., extrabudgetary funds) there is no immediate trade-off for other government expenses but this option may reduce the future state budget and may impact the citizens with some delay. Since the subsidies result in increased purchasing power, a short-term rebound effect can be expected.

To determine the total effect, the rebound and the prebound effect should be added up which would however require a reliable quantitative analysis.

4.1 Putting the results into the perspective of earlier work and of recent developments

The following brief review summarises the current state of knowledge regarding the effectiveness of policies targeting renewable heating systems and related policy objectives. As is widely known and in line with the choice made for promoting renewable heating in Switzerland, challenging policy objectives are typically addressed by a mix of policy measures, allowing to increase the overall effectiveness and accounting for the presence of pre-existing measures [49,50]. For example, the Swiss cantons implemented their measures in the context of a CO₂ levy which amounted to 96 CHF/t CO2 from 1.1.2018 until 31.12.2021 and which was raised to 120 CHF/t CO₂ on 1.1.2022 but which had not triggered a turnaround towards decarbonisation of heating.

As Steg, Dreijerink, & Abrahamse [51] emphasized, public acceptance is a crucial prerequisite for the successful implementation of energy programs, in particular in cases of voluntary measures. Energy policies are difficult to accomplish without public support due to opposition and politicians’ unwillingness to carry out programs that lack popular support (ibid.). The support depends to a great extent on psychological factors. Policies are rated as less acceptable when they are perceived as unjust, when they threaten people’s right to free choice, and when they are thought to be ineffective in addressing the environmental issues at hand (e.g., [51]). In addition, the studies demonstrate that individuals evaluate policies to be more acceptable when they are more aware of the environmental issues, feel more accountable for them, and have a larger moral commitment to help solve these challenges [51].

The use of incentives (rewards) or disincentives (penalties) to alter behavior is one of the key characteristics of energy policies (e.g., [52]). “Pull” measures involve financial incentives and knowledge dissemination, and are intended to promote the desired behaviour [51]. Since they are more likely to be viewed as ‘voluntary’, they tend to result in favorable attitudes, according to Geller [52].

On the other hand, coercive policies may force people to modify their behavior and may therefore appear to be more effective than the so-called pull measures, which are noncommittal in nature. However, coercive measures are frequently accompanied by negative affect, sentiments, or attitudes toward the behavior change since they are more likely to jeopardise individual freedom. According to some experts, this makes them potentially less successful than pull measures [52]. For example, in a study conducted by Contzen, Handreke, Perlaviciute, & Steg [53], researchers measured people’s emotional responses to two policy alternatives that would increase the use of heat pumps in a neighborhood of the city Groningen in the Netherlands. In accordance with the reactance theory [54], the policy option that mandated the heat pump installation was perceived as threatening people’s freedom more and, as a result, produced stronger negative and weaker positive feelings than the alternative, which encouraged voluntary adoption (ibid.). Importantly, the attitudes about policy choices were also influenced by people’s values. Environmentally oriented values were solely associated with stronger good feelings in relation to the policy option supporting voluntary adoption (but not towards the mandatory policy), whereas stronger egoistic values were associated with stronger negative emotions, especially toward the option requiring the adoption of heat pumps (ibid.).

In line with the above, Stadelmann-Steffen et al. [55] as well as Kammermann and Ingold [50] found for Switzerland that the stakeholders (policymakers, administration, economy, civil society) prefer soft policy measures, in particular information and advice but also subsidies and tax reliefs, to regulatory measures. Coercive measures may nevertheless be preferred when constraints call for them, e.g., related to environmental or landscape protection, leading to local (cantonal and municipal) master plans which define the principles. Coercive measures may also be chosen in the case of alignment with individual values and convictions [56]. As already pointed out above, it is, however, a prerequisite for the support of policy measures by citizens, that these are convinced by their effectiveness [30]. In addition, a high risk aversion and high investment aversion was identified for Switzerland which calls for politics of small steps and the application of business models that allow to convert upfront investment into operating cost (e.g., contracting). The chosen solutions must not be too expensive and they must have no or very limited disadvantages for incumbents. Successful Swiss policies imply effective communication and awareness raising as well as active involvement and participation of the stakeholders (co-design, co-decision), typically over long periods of time and engaging the most affected. Furthermore, Freyre et al. [45] revealed that, in order to increase the effectiveness of renewable heating programs, the support for individuals should not only include financial instruments; additionally, partnership initiatives with local installers should be introduced to help homeowners select competent technicians.

It is noteworthy that, in spite of being fiercely contested policy area, there have been only a limited number of recent studies on the effectiveness and acceptance of decarbonisation policies for buildings. Smith et al. [57] point out that “public attitudes towards decarbonised heating remain under-researched and poorly understood”. In line with the present study, they find that the support of subsidies for installing heat pumps is significantly higher than the support of bans of installing new oil and gas boilers. They find that a too extensive restriction of the freedom of choice can lead to strong public response. Along the same lines, Salite et al. [58] conclude from a survey in socioeconomically deprived neighbourhoods that the adoption of sustainable heating systems should remain optional and that not only financial incentives will be needed to promote renewable heating adoption but also non-financial measures (e.g., information and awareness raising). Bosetti et al. [59] summarize examples of backlash to climate policy which they put into the context of right-wing populism; prominent examples can be found in the U.S. (see, e.g., [60]). Edenhofer et al. [61] acknowledge that bans on natural gas-based heating have encountered strong opposition by homeowners who feel that their unique circumstances are ignored. On this basis they argue in favour of targeted bans based on thorough analysis.

Today, Nordic countries (in particular, Norway, Finland and Sweden) have by far the highest diffusion levels of heat pumps in Europe (40%-60% of all buildings according to [62] compared to 23% in Switzerland; [63]). According to Lowes et al. [64] this can be explained by a the implementation of a set of policy measures (referred to as “heat pump policy package”, consisting of i) coordination and communication ii) economic or market instruments iii) financial support and iv) regulations and standards. On the other hand, “single policies on their own are unlikely to drive heat pump deployment at the levels required by global decarbonisation goals” [64]. This is confirmed by Johansson [65] who explains that rapid expansion of heat pump implementation in Sweden was preceded by two decades of industrial and entrepreneurial activity, academic research and interventions by electric utilities and governmental agencies. Similarly, while Norway banned the use of oil and gas boilers in new buildings and major renovations in 2016, this was anticipated by a long history of subsidies for heat pumps (in 2003), a carbon tax (in 1991) and a reduced tax on electricity used for heating (in 2017; [66]).

In summer 2025, new cantonal model regulations were released in Switzerland (CMR2025). While its predecessor (CMR2014) had required a renewable energy share of at least 10% for new heating systems, CMR2025 calls for the exclusive use of renewable energy or of waste heat. Since all cantonal energy directors had to agree to the new model regulations, this is a remarkable achievement, indicating that the strategy of gradually increasing the required renewable energy share while preparing the ground and ensuring sufficiently high acceptance has been successful. This indicates that it is possible to implement more coercive policies while ensuring acceptance.

4.2 Limitations and further reflections

The lack of official data on actual implementation of heat pumps at this stage is a drawback, making it impossible to state with certainty the effectiveness of the discussed policy measures. Similarly, data on investment costs, energy costs, interest rates etc. are subject to significant uncertainties, especially during more recent periods.

Most retrofit experience has been made with heat pumps in single-family houses, while some theoretical and real-life case studies show that the implementation in existing multi-family buildings (especially larger and older ones) can be challenging, potentially calling for hybrid solutions [67,43]. This should be considered in any policy design.

As mentioned in section 2, interviewees were chosen from cantons that are known for their proactive stance in energy and climate policy. This is likely to have led to a more optimistic feedback compared to the inclusion of cantons which are lagging behind on these matters. On the other hand, the recent implementation of CMR2025 implying the ban of fossil boiler replacement shows that the proactive cantons managed to convince the laggards within a rather short period of time (between 2022 and 2025). However, it remains to be seen in the coming years how easily the cantons will be able to implement these policies and how the building owners and the tenants will subsequently perceive this change.

As further limitation, the number of interviewees was relatively low. In particular, the involvement of further housing companies and tenant associations may have allowed to obtain a more comprehensive understanding. Future research should therefore ensure a more comprehensive sample of stakeholders.

Local policymakers aim to make use of local resources, with bioenergy being a promoted option in some rural areas (e.g., Jura). While this strategy is understandable, heat pumps should, in general, be prioritised in order to save bioenergy resources for more demanding applications than space heat [68]. Further analysis is required to provide more rigorous scientific proof for the need of a merit order among the renewables used for heating.

The implementation of coercive policies can trigger controversies and partly unwanted reactions. According to the experience made in Switzerland and Germany, a share of owners decided to install a new fossil fuel boiler just before the regulation on phase-out entered in force. This decision may be enhanced by special offers and campaigns run by installers. It may be worthwhile to investigate how this phenomenon can be mitigated. One of the countermeasures is to build more trust by ensuring the citizens that nobody will be left behind, i.e., that specific measures will be taken to address modest and disadvantaged households (owners and tenants). Clarity about the scope should be ensured by the development of local heating (and cooling) plans, allowing owners, installers and the respective industries (installers, district heating and cooling companies, distribution grid operators etc.) to prepare for future requirements.

The calculations presented in section 3.4 are simplified by referring to only one building with single input values for each of the parameters. This type of analysis is nevertheless helpful by allowing to obtain a first understanding of the interrelations. In comparison, Li et al. [33] had access to a more comprehensive dataset, including data on investment costs and subsidies for individual installations, allowing to consider the value ranges. However, this information was only available for two cantons (out of 26 cantons in total) and not for all parameters (e.g., no data were available on hourly heat demand, efficiency, operational cost etc.). In conclusion, the type of analysis presented in the present paper should be reconducted once access can be obtained to larger real-life datasets. This will allow to draw more robust conclusions.