Introduction

In climate change mitigation research, demand-side options are increasingly considered as a crucial supplement to technology-focused supply-side measures. They can serve to reduce the reliance on novel but potentially risky carbon dioxide removal methods in climate stabilization scenarios consistent with the Paris Agreement1,2,3,4 as well as to mitigate the material impact of energy transitions5. However, despite surging research highlighting major mitigation potentials of demand-side options6,7, systematic analyses of actually realized or implemented demand-side mitigation policy have remained scarce8. As a result, the current status of demand-side climate policy-making remains poorly understood, obstructing assessment and exploration of effective demand-side mitigation strategies. Hence, there is a need to investigate existing demand-side climate policy mixes, their evolution, composition and foci, in order to improve climate-policy making as well as scenario development2,9.

Demand-side options refer to policies and measures designed to reduce material and energy consumption, along with associated greenhouse gas (GHG) emissions, by altering final demand for goods and services, without compromising or potentially improving well-being10. These options typically target adoption and choices of technologies in consumption, consumption and lifestyle patterns, as well as related systems of service provisioning and interconnected production-consumption infrastructures2,10.

Demand-side mitigation options can be categorized into improve, shift and avoid measures8,10. Improve measures aim at enhancing the material and energy efficiency of existing technologies, modes of service provisioning, and end user adoption, such as driving less emission-intensive cars. Shift measures promote the switch to modes of service provisioning that are less emission-intensive per unit of service provided, e.g., using public transport. Avoid measures focus on avoiding demand for certain goods and services altogether, by redesigning service provisioning systems to reduce unnecessary resource consumption, e.g., by reducing distances travelled. While many consumption choices are ultimately made by individuals or households, these choices are strongly influenced by social norms and political decisions at various political levels that shape available infrastructure (e.g., public transport availability, quality, and costs), economic incentives, and regulatory frameworks11,12,13.

Approximately two-thirds of global GHG emissions are directly or indirectly linked to household consumption14,15,16,17, with the remainder attributed to emissions from government infrastructures (e.g., hospitals) and capital investments. Therefore, literature on demand-side options highlights major untapped mitigation potential, particularly if a comprehensive and targeted policy mix encompassing improve, shift, and avoid measures is implemented8,11,18. The most effective mix of improve, shift, and avoid measures is suggested to vary across sectors. High improve and shift potentials have been identified in the housing sector (e.g., refurbishment and renovation, shift to renewable energy), while high shift and avoid potentials are found in the transport sector (e.g., shift to public transport, avoid long-haul aviation, promote short-distance urban structures)11,15. However, while most improve measures are deemed more politically feasible, they are also susceptible to being undermined by rebound effects8,19,20. In contrast, many shifts and, particularly, avoid measures are considered unpopular and risky, as they often involve breaking entrenched routines or disrupting dominant modes of living11,21,22.

However, despite the prominence of demand-side options in the recent IPCC report23, relatively few studies have systematically analyzed the evolution and status of actual demand-side policy instruments and mixes using the avoid-shift-improve framework (henceforth ASI-framework). Most existing studies have focused on the application of the ASI-framework to the transport sector, where it originated8,24. Research has emphasized the significance of experimentation for developing niches for ASI-based measures in cities to support low-carbon mobility transitions25, assessed the potential of reducing car emissions in urban areas26 and mobility services for sustainable transport system transformation27 based on the ASI-framework, and explored potential impacts of various ASI-options in the transport sector on energy systems28.

Most recently, the ASI-framework has also been adapted to study energy policy mixes8, demonstrating that most policies focus on improving efficiency of current technology and shifting to low-carbon alternatives, while policies prioritizing energy demand reduction are underrepresented (amounting to merely 6–22% of policies). Other studies have focused on climate and energy policy mixes targeting household consumption (i.e., the demand-side)14,20,29, without, however, consistently classifying policies along the ASI framework. In line with tangential research on sufficiency in energy systems30,31, these studies indicate that demand-side oriented policy mixes have been skewed towards efficiency measures while neglecting avoid or sufficiency options8,11,14,20. However, to what extent these initial findings on the asymmetric composition of demand-side mitigation policy correspond to the overall evolution and composition of demand-side policy mixes across sectors and policy levels has not been systematically investigated.

Meanwhile, research on climate and energy transition policy has increasingly paid attention to cross-sectoral and multilevel policy mixes and their characteristics, such as coherence, comprehensiveness, and synergies, as well as temporal dynamics and patching18,32,33,34,35,36. Nonetheless, much of this work has primarily focused on supply-side measures, leaving demand-side mitigation policy mixes and methodologies for assessing them largely unexplored37. Advancing demand-side mitigation options, therefore, requires a deeper understanding of the policy mixes across sectors and policy levels currently in place, and how they evolved.

Here, we provide a longitudinal, multilevel analysis of the evolution, composition, and foci of demand-side climate change mitigation policies for the case of Austria from 1995 to 2024 in the transport and housing sectors. As in other OECD countries14, these two sectors are responsible for the majority of Austria’s household GHG footprint emissions (60.8% in 2022, Fig. 1), which in turn account for the majority of total footprint emissions and fall only slightly short of total production-based emissions (Fig. 2). Our sample includes policies and major public investment decisions aimed at influencing investment or consumption choices of households, or households’ behavior, in ways that reduce household GHG emissions (see Method section for details on our classification and coding approach). We systematically coded, mapped and classified policies according to policy target, instrument type, and the demand-side mitigation option pursued, discerning improve, shift and avoid measures based on the established ASI-framework (Table 1).

Fig. 1: Product group composition of the GHG footprint of Austrian households for 1995, 2000, 2005, 2010, 2015, 2020, and 2022.
figure 1

Calculations are based on input-output tables of EXIOBASEv.3.8.276, enhanced with Austrian production-based GHG emissions data from the Environment Agency Austria82 (see Methods). Indirect housing GHG emissions comprise electricity, residential buildings operating services, district heating, production of heating fuels, materials for housing maintenance etc. Direct housing GHG emissions result from burning of heating oil, gas and coal in private households. Indirect mobility GHG emissions include production of fuels, travels (flights, other transport modes like bus or train), production of cars, vehicle spare parts, maintenance etc. Direct mobility emissions result from burning of diesel and gasoline in private cars.

Full size image
Fig. 2: Austrian production and differentiated consumption-based GHG emissions, 1995–2022.
figure 2

Consumption-based GHG emissions are based on own calculations using input-output tables of EXIOBASEv.3.8.276; production-based emissions are from the Environment Agency Austria82 (see Method section).

Full size image
Table 1 Demand-side mitigation policy classification based on Creutzig et al.10
Full size table

Focusing on a single country allowed us to zoom in on the multilevel nature of demand-side policy mixes and their evolution over time, encompassing the EU, national, and provincial levels (selecting Austria’s two most populous provinces Vienna and Lower Austria). This approach enabled us to investigate how demand-side policy mixes vary across policy levels based on a high-resolution policy database. In contrast to large cross-national climate policy mix assessments that rely on established databases, such as the OECD Climate Policy Databasee.g.,38, which lists merely 12 policies for Austria between 1995 and 2024, our database includes a total of 184 federal-level policies for the same period. By additionally incorporating EU and provincial-level policies, we created a sample of 236 demand-side climate policies in the transport sector and 120 in the housing sector, respectively.

Austria serves as an interesting case study: The country adopted highly ambitious climate targets (net-zero by 2040) but did not achieve substantial emission reduction until very recently (2022). It thus lags behind EU average in terms of achieved mitigation outcomes, which has been mainly attributed to a lack of policy integration39,40, as well as political deadlocks due to strong federalist41 and corporatist structures42. We therefore consider Austria to be a particularly relevant case study, as it features a relatively broad multi-level portfolio of climate policies while simultaneously enabling a closer analysis of the gaps and deficiencies of existing policy mixes.

Our study provides evidence on what demand-side policies are already represented in the transport and housing sector across different policy levels over time. This allows us to identify relevant gaps in sectoral demand-side policy mixes and to discuss our findings against the background of newly available data on carbon inequality trends in Austria from 2000 to 202043, along with broader political implications. A limitation of this study is that merely assessing the quantity of policies across policy types and instruments does not account for the fact that individual policies vary in scope, effect, and importance. We compensated for this by complementing policy mapping with semi-structured expert interviews and an expert survey among peer researchers (see methods section for details). Nevertheless, further research is needed to explore the effectiveness of individual demand-side policies and policy mixes, as well as to identify associated leverage points for demand-side mitigation44,45.

Results: demand-side policy mixes exist but remain partial

Among the 356 demand-side mitigation policies identified in the transport and housing sectors, 218 are shift measures and 121 are improve measures. By contrast, avoid options are largely neglected in both sectors (17 measures). In the transport sector, shift measures constitute 64.4%, improve measures 31.7%, and avoid measure only 3.8%. Similarly, in the housing sector, shift measures account for 55%, improve measures for 38.3%, and avoid measures for 6.7% (Fig. 3).

Fig. 3: Identified demand-side policies in the transport and housing sector.
figure 3

Policies are categorized based on the avoid-shift-improve framework and policy target addressed (upper part), and policy instrument type for different demand-side options and the total policy mix (lower part). The underlying policy dataset is provided in the source data (see Data Availability Statement).

Full size image

Transport sector: targeting cars and modal shifts, neglecting avoid options and air travel

In the transport sector, private cars (including battery electric vehicles) were the most frequently targeted focus of policies, accounting for 47.4% of all measures. Most improve policies aimed to reduce the attractiveness of driving or buying cars with large or inefficient internal combustion engines. These policies, particularly prevalent in the policy mix in the period from 1995 to 2010, increased fuel taxes, introduced tax benefits for low-CO2-emission cars (as part of an engine capacity-related insurance tax), and integrated a fuel combustion and CO2-emission-based component into the motor vehicle registration tax (Fig. 4). Measures that increased the cost of driving internal combustion engine cars may have indirectly encouraged households to adopt alternative transport modes. However, these were categorized as improve measures, as their primary objective was to incentivize the purchase of more fuel-efficient cars—at least for households that could afford them. Especially since the mid-2010s, additional improve measures directly targeted the adoption of battery electric vehicles. These measures have predominantly relied on price premiums, tax exemptions, investments in the deployment and accessibility of charging infrastructure, and other forms of regulatory privileging, such as designated parking spaces.

Fig. 4: Multi-level demand-side policy mix and development of household GHG footprint emissions in the transport sector.
figure 4

The upper part presents selected demand-side policies. Note that many policies included in the policy dataset are contained in strategy documents and are not displayed individually here. Strategies containing multiple demand-side measures are shown in light beige, i.e., not assigned to the avoid, shift, or improve color code. The lower part shows the development of the household GHG footprint emission in the transport sector, including direct and indirect GHG emissions. Data are based on own calculations using EXIOBASEv.3.8.2 and data from the Environment Agency Austria.

Full size image

Shift measures targeting private cars predominantly aimed at disincentivizing their use, rather than promote the adoption of more efficient vehicles, thus encouraging a transition to alternative transport modes. A key example of these measures is the introduction of short-term parking zones and parking fees in Vienna in 1995, followed by their gradual expansion, which increased the attractiveness of other transport options for inner-city mobility. Beyond discouraging private car use, policies aimed at promoting shifts toward other transport modes—primarily public transport and cycling, but also walking and car-sharing—constitute a substantial proportion of demand-side measures in the transport sector (Fig. 3). In Vienna, a distinct policy pattern emerged combining push measures such as parking zones and fees, with pull measures encouraging shifts to alternative transport modes46. These pull measures include investments in public transport infrastructure, cycling and park-and-ride facilities, subsidies for cargo e-bikes, and the introduction of the highly affordable 365-Euro annual public transport ticket. Moreover, the concept of “City of Short Distances” (Stadt der kurzen Wege) has been successively integrated as a key principle into urban planning in Vienna, with a specific focus on avoiding unnecessary motorized travel through compact structures and functional mixes (residential, commercial, recreational, and educational) in existing and new districts. However, while successful in limiting urban sprawl and suburbanization within the city boundaries47, work-related commuter flows from other federal states to Vienna have continuously increased48,49.

We identified relatively few policies aimed at promoting train travel beyond urban public transport (Fig. 3). These include substantial investment plans to expand and enhance rail infrastructure and the introduction of the KlimaTicket in 2021, a yearly flat-rate ticket covering all rail and public transport services. However, many policies explicitly disincentivizing private car use may also implicitly encourage increased train travel. Stunningly, we only found four demand-side climate policies addressing air travel.

Overall, the transport sector exhibits a relatively extensive policy mix emphasizing shift measures, aimed at encouraging individuals to transition away from private car use toward alternative modes of transport. Beyond the growing focus within improve measures on promoting the adoption of battery electric vehicles, we could not identify substantial changes in the focus of the policy mix over time. Simultaneously, certain areas (notably air travel) and policy options (particularly avoid measures) have remained largely neglected. Except for Vienna’s focus on compact city structures in urban planning, the avoid measures identified were predominantly limited to declarations of intent to explore options for reducing mobility needs through spatial planning and promoting teleworking.

Housing sector: improving efficiency, shifting to renewables, omitting avoid potentials

Improve and particularly shift measures also dominate the policy mix in the housing sector (Fig. 3). The strong presence of shift measures is largely attributable to federal and provincial subsidies for photovoltaic systems, solar thermal energy, heat pumps, and biomass-based heating systems, introduced in the mid- to late 2000s (Fig. 5). These subsidies were typically renewed annually over extended periods and were complemented by bans on fossil fuel-based heating systems at both federal and provincial levels, initially targeting coal and oil, and more recently, natural gas. Additional shift measures include information campaigns and regulatory instruments, such as building standards and the suspension of permit requirements, to facilitate the transition to renewable energy and heating options.

Fig. 5: Multi-level demand-side policy mix and development of household GHG footprint emissions in the housing sector.
figure 5

The upper part presents selected demand-side policies. Note that many policies included in the policy dataset are contained in strategy documents and are not displayed individually here. Strategies containing multiple demand-side measures are shown in light grey, i.e., not assigned to the avoid, shift, or improve color code. The lower part shows the development of the household GHG footprint emission in the housing sector, including direct and indirect GHG emissions. Data are based on own calculations using EXIOBASEv.3.8.2 and data from the Environment Agency Austria.

Full size image

Improve policies primarily targeted on enhancing the energy efficiency of buildings. Central to these efforts were federal and EU building and construction regulations, alongside the mandatory provision of energy performance certificates for houses and apartments. A smaller subset of improve measures aimed at directing household demand towards more efficient household appliances, mostly through information campaigns.

The few avoid measures identified primarily aimed to raise public awareness of energy-saving practices or to promote passive houses through regulatory standards (e.g., introducing quality standard labels) and subsidies for buildings with very low energy consumption indicators. Notably, these measures aimed at avoiding energy consumption rather than GHG emissions associated with carbon-intensive construction materials. Only since 2022 were additional avoid measures introduced at the EU and federal levels to encourage the repair of goods using regulatory and economic instruments. Remarkably, no avoid measures were found that directly addressed the avoidance of new construction.

Discussion

Our analysis reveals that actually implemented demand-side measures have remained partial, with a notable neglect of avoid options. Although the Green Party’s first participation in the national government led to a surge in demand-side policymaking and subsidy volumes, accompanied by relevant reductions in production-based GHG emissions (−5.8% in 2022, −6.4% projected for 2023), the dominant policy focus remained on shift and improve measures persisted, alongside continuously high household GHG footprint emissions (Fig. 2).

Regarding the multi-level and longitudinal dimensions of analysis, policy data and evidence from expert interviews indicate that EU policies, specifically binding national emission reduction targets and associated sanctions as part of the EU Effort Sharing Decision (2009), were often decisive for the decision to introduce demand-side measures in Austria, especially subsidy schemes in the housing sector on the national and provincial level. This corresponds with findings that most regulations on low-emission and sustainable buildings have been implemented due to EU legislation50. In particular, our interview evidence suggests that subsidies promoting demand-side shifts allowed policy-makers to signal commitment to achieving EU targets without having to face the challenge of implementing potentially contested regulatory interventions within Austria’s consensus-oriented corporatist political setting. At the same time, these subsidies frequently faced budgetary constraints and downsizing, except for the period 2019–2024 when reinforced EU and national climate ambitions resulted in a substantial expansion of public expenditure in this area (drastically scaled-back, however, in early 2025 due to Austria’s fiscal crisis).

Sectoral trends in household GHG footprint emissions (Figs. 4 and 5) suggest that the lack of avoid measures has been particularly impactful in the transport sector. Despite notable reductions in the mid-2000s and 2010s, mobility-related household emissions in 2022 remain higher (17.6%) than in 1995. By contrast, housing-related household footprint emissions have experienced a decline, falling well below the 1995 level (−29.4%). While direct attribution of emission reductions to specific policies is beyond the scope of this paper, these trends indicate that shift measures, which incentivize and enable households to transition to renewable energy, along with improve measures aimed at increasing buildings energy efficiency, have been relatively effective, despite counterproductive subsidies (specifically an industry-funded subsidy for oil-fired heating systems). Direct housing-related emissions (i.e., fuel use) correspondingly saw the largest reduction (33.6%), considerably lowering their relative contribution to overall household footprint emissions43 (from 13.1% in 1995 to 9.3% in 2022) (Fig. 1). This suggests that Austria’s subsidy-heavy demand-side policy instrument mix was modestly successful in reducing emissions in the housing sector—mostly those of home owners—where mitigation is particularly capital-intensive. At the same time, it remained largely unsuccessful in the mobility sector where deeper interventions into behavioral patterns and public infrastructure investment are required.

Correspondingly, mobility-related emissions have remained the largest contributor to household GHG emissions, with their share in total footprint emissions increasing from 26.5% in 1995 to 33% in 2022 (Fig. 1). This contrasts with the higher density of demand-side policies in the transport sector compared to the housing sector (236 as opposed to 120 policies in housing). A variety of policies has been implemented that make driving fuel-inefficient cars more expensive, incentivize the adoption of battery electric vehicles, and encourage shifts to alternative transport modes. However, these measures appear to have been partially undermined by insufficient policy integration, including counterproductive subsidies such as the design of the commuter allowances, the diesel tax privilege, and tax benefits for company cars39,40. Aside from Vienna’s commitment to promote compact city structures through spatial planning, the federalist division of competencies in spatial planning has also promoted urban sprawl, landscape fragmentation and related mobility-related household emissions51. Moreover, our findings suggest that household demand for air travel has remained hardly addressed by existing demand-side policies, echoing results from other country-focused studies52. Overall, stubbornly high mobility-related household emissions highlight the urgent need to exploit the mitigation potentials of avoid measures identified for this sector6,15.

The notable above-average income elasticity of mobility demand highlighted in research on carbon inequality in high-income countries, including Austria43,53,54, and the lack of demand-side policies addressing air travel, suggests that avoid measures could unleash special potential when targeting mobility patterns of high-income groups, e.g., through levies on frequent flying or luxury-focused taxation55, or a ban of short-haul flights or private jet flights. Such a focus could increase the political feasibility of avoid measures, often deemed unpopular or risky11,56, as their costs remain concentrated on a relatively small subset of high-emitting households, and particularly a ban of private jet flights enjoys broad support across party affiliations in several EU countries57. Concomitantly, other avoid options increasing well-being or providing co-benefits by lowering mobility needs through urban design or teleworking, could target a wider spectrum of households across various income levels10,24,58.

While continuously high mobility-related emissions underscore the urgent need for avoid measures in the transport sector, our findings also suggest that the potential for avoid options in the housing sector have not yet been leveraged59,60. For example, policies could incentivize and support the adjustment of household to dwelling size through relocation grants, for instance after children have moved outcf.61. Additionally, strategies that increase political feasibility of avoid options by selectively targeting high-income lifestyles are conceivable, e.g., limiting excessive floor space per capita as a major contributor to household GHG emissions43, or avoiding new construction and the associated use of carbon intensive-building materials62 by mobilizing existing housing stock through effective vacancy fees.

While infrastructure investment constitutes a relevant portion of shift measures, particularly in the transport sector, our findings also show that both improve and shift measures heavily rely on economic instruments (Fig. 3). Worryingly in light of growing concerns over popular backlash to climate policy63,64,65, however, the composition of these measures suggests that they tend to make emission intensive practices (e.g., driving fuel-inefficient cars) more expensive, but access to subsidy schemes promoting the adoption of low-carbon alternatives (e.g., battery electric vehicles, renewable energy heating systems in privately owned homes) has tended to be biased towards high-income households that can afford such investments66. While this may be less problematic in urban areas with existing infrastructure that facilitate modal shifts, as evidenced here by Vienna, it is specifically concerning with respect to low-income households in rural areas, who face rising costs but lack enabling conditions for modal shifts. Our findings, therefore, also indicate an urgent need to make demand-side subsidy schemes more inclusive, ensuring broader access for low- and medium-income households.

Conclusions

While a surging body of literature points to the potentials of demand-side option for climate change mitigation6,11, our case study shows that currently existing demand-side policy mixes remain partial, particularly lacking avoid options. A notable exception is Vienna’s commitment to the concept to the “City of Short Distances” in spatial planning. Covering the evolution of demand-side policy across different sectors and policy levels in Austria over almost 30 years, our study essentially reflects similar findings from other European countries that avoid or sufficiency measures have remained underrepresented in mitigation policies addressing household GHG emissions14,20,30. Nevertheless, while generally in line with earlier findings from other European contexts, some country-specific factors likely account for the stark absence of avoid options in our case: Given Austria’s comparatively strong fiscal position, pressure to achieve binding EU mitigation targets prompted policy-makers to implement subsidy schemes focussed on shift and improve, allowing them to circumvent political resistance against regulatory or even avoid-focussed policies imposing visible costs, particularly on corporate actors (e.g., the construction industry if new building construction is avoided, or on the energy utility sector if energy demand decreases). In addition, spatial planning competence distribution favoring settlement dispersion are at odds with policy efforts to avoid mobility, at least outside of Vienna.

These country-specific aspects notwithstanding, key factors explaining the lack of avoid measures in demand-side policy mixes are evidently more generic and also apply to other European countries and beyond. Literature on demand-side solutions specifically highlights concerns over controversiality and popularity of such measures, especially given political sensitivities surrounding governmental initiatives to alter individual behavior11,20. Political economy factors, such as vested interests, can be considered at least equally important6, as measures avoiding demand tend to frustrate expectations in growth, profit and investment opportunities on the supply side8, including the renewable and cleantech sector30. However, more research is needed to better understand what factors determine the popularity and political feasibility of avoid measure (and policy-makers perception thereof)8, especially given current climate policy backlash63,64. This is particularly crucial, as emerging contributions on political strategies for climate policy56 highlight broad support for supply-side measures like green industrial policy subsidies, but have largely overlooked demand-side optionscf.57.

To strengthen avoid measures as part of comprehensive demand-side policy mixes, stressing positive effects on human well-being10 and participation11 is essential, as well as dissociating avoid measures from generalizing notions of renunciation, except when upper limits specifically target high-income, emission-intensive lifestyles67,68,69. Equally important are novel findings on public support for sufficiency or degrowth measures70,71 and how such support can be actively shaped72. Ultimately, however, advancing urgently needed avoid options—despite potential political opposition—will require political strategies which consider actor constellations and how they can be leveraged and transformed in multilevel politics and governance settings. Relevant examples include recruiting allies through strategic stakeholder engagement (e.g., local business that benefit from compact mobility patterns), expanding the set of policy winners through issue linkage (e.g., linking traffic avoidance with work-time reduction as part of a policy-package for stress reduction), or limiting the access of vested interests that act as policy opponents (e.g., more systematically isolating spatial planning bodies from the influence of the construction industry)56,73.

Methods

Method 1: coding and classifying multilevel demand-side policy mixes

The demand-side policy mapping includes policies and public infrastructure investment decisions from 1995 to 2024 (excluding December 2024) that aimed at influencing investment or consumption decisions or behavior of households, in ways that reduce household GHG emissions in the transport and housing sector. While household GHG footprints and demand-side climate policy analyses have been linked for a given moment in time14,20, we here approach demand-side policy mixes from a longitudinal perspective. This allowed us to analyze the evolution of demand-side policy against the background of a 27-year time series of household GHG footprints. While investigating the effectiveness of individual policies in terms of GHG emission reductions is beyond the scope of this paper, this approach yields relevant insights into how household GHG emissions and demand-side mitigation policy co-evolved, providing the basis for further research into causal dynamics. 1995 was chosen as a starting point as EXIOBASE data is available from this year onwards, and also coincides with the emergence of climate policy as a dedicated policy area in the EU in the mid-1990s74.

We adopted a multilevel perspective, focusing on policies at the EU level (regulations, directives, strategies and action plans), the federal Austrian level as well as the level of two Austrian provinces (Vienna and Lower Austria). We considered the provincial level because many competencies, particularly in the areas of transport and housing, are located there75. Vienna and Lower Austria, Austria’s two most populous provinces, were selected for their contrasting profiles in terms of homeownership and tenancy rates, rural or peri-urban vs. urban settlement structures, and related mobility patterns. We focused on the housing and transport sector because these two sectors—as in other OECD countries14—account for the majority of household GHG footprint emissions (Fig. 1).

We identified demand-side policies and infrastructure investments based on multiple research strategies. First, we conducted a systematic literature review on relevant climate policy in Austria based keyword search in Scopus and Google Scholar (keywords: “Austria” AND (“transport” OR “mobility” OR “housing”) AND (“policy” OR “policies”)) between January and October 2023. This reaped a result of 199 articles and book chapters (109 transport, 90 housing), 90 of which were selected for systematic coding after title, abstract and keyword screening. Second, literature review was followed by a targeted web search to locate the specific policies mentioned in the literature, such as legal texts or policy documents. An additional web search was conducted based on the literature review to investigate when a particular subsidy schemes or program mentioned for a specific year or time span had been implemented, and when it expired.

Third, we systematically screened websites of relevant federal ministries and provincial government agencies (particularly based on systematic screening of press releases for the available years) as well as websites of initiatives or institutions in charge of implementing climate policy measures on behalf of the federal or provincial administrations (e.g., the initiative “klimaaktiv” implementing various information campaigns and subsidy schemes on behalf of the federal government).

Fourth, many policy documents mentioned previous policies in the area which they built upon or which they replaced or modified, which allowed us to trace policy evolution backwards. Fifth, to triangulate the results from these search strategies, we conducted four semi-structured expert interviews with civil servants at the Federal Ministry of Climate Action, Environment, Energy, Mobility, Innovation and Technology as well as the Director’s Office for Climate Affairs at the city of Vienna.

The first two interviews with civil servants at the federal ministry were conducted in March and April 2023, with a duration between 45 and 60 min. The interview guidelines’ focus was on validating the policy data corpus and our coding of it (the relevant datasets were sent to the interview partners), identifying remaining gaps in the data and avenues for further policy search, and on obtaining their opinion on whether certain demand-side policies were particularly relevant or effective in terms of emission reduction. We also inquired whether they believed certain policies or policy mixes have been weakened due to insufficient policy integration, where they saw unaddressed areas or gaps in the implemented policy mixes, and which results of the policy analysis as well as the analysis of household GHG footprint development in Austria (as part of the overarching research project this study is part of) they considered particularly relevant or helpful for their work.

One interview partner was specifically selected because of the long employment duration at the ministry in relevant functions, covering almost the entire study period. Nevertheless, given the longitudinal character of the study, we assumed that the ability to recall events over the entire period—particularly the early phase—would be limited. Therefore, these expert interviews primarily served to validate findings, explore relevant moments of change or breaks as well as multi-level policy-making dynamics, and identify further research directions. Two shorter interviews of ~20 min with a civil servant at the federal ministry and at Vienna’s climate office were conducted in December 2024 to validate the final policy data corpus as well as the coding of policies adopted towards the end of the study period after the first round of interviews. The interviews were transcribed (unless interviewees declined to be recorded) and coded for gaps in or inconsistencies with the dataset, as well as with regard to the relevance of individual policies, gaps in the policy mix, relevant breaks or moments of change, and missing policy integration.

Sixth, we also presented key intermediate findings at a stakeholder workshop with federal and provincial climate policy makers in Austria as well as NGOs involved in climate policy in November 2024. While the main aim of this workshop was to identify ways to ensure or increase social acceptance and political feasibility of avoid or sufficiency measures in demand-side climate policy-making (the focus of a related module of the overarching research project), it also gave us the opportunity to present the preliminary findings of this study on the evolution and composition of existing demand-side policy mixes and obtain feedback on the comprehensiveness and consistency of our policy data. Seventh, we shared and discussed our policy dataset with members of the Austrian climate policy research community in two online meetings held in July and December 2024 to obtain peer-feedback and pinpoint gaps in the dataset.

This resulted in a dataset comprising 356 policies in total (236 transport, 120 housing). In parallel with data collection, policies were systematically coded along three dimensions: (1) policy target (2) type of policy intervention, and (3) demand-side mitigation option pursued (see Table 2).

Table 2 Three-dimensional policy coding scheme
Full size table

While assigning codes to policies was straightforward in most cases, the following coding rules were developed to guide the coding exercise in intricate cases and cross-checked for consistency by at least three authors familiar with the dataset and respective coding rules.

  • If policy measures aimed at multiple targets or relied on several policy instruments, they were coded multiple times to adequately reflect the diversity of the policy mix.

  • Subsidy schemes (e.g., purchasing premia for battery electric vehicles) were counted as a single policy measure even if they were renewed with minor adaptations over several consecutive years. However, if subsidy schemes were substantially expanded in scope or volume (e.g., more than doubled), we coded them as an additional policy measure. This approach enabled us to more accurately determine the primary focus areas of existing policy mixes. Counting subsidy schemes on an annual basis would have created a strong bias, making policy mixes appear disproportionately focused on subsidies.

  • Similarly, information campaigns (e.g., on energy saving) were counted as a single policy even if they were relaunched multiple times or relevant information was provided over several consecutive years.

  • Policy measures were considered to target “rail” instead of “public transport” if the primary focus was on longer-distance intercity connections. By contrast, they were coded as targeting public transport if the focus was on public transport within cities or between cities and surrounding areas.

  • Policies targeting the use of private cars fall into two broad categories: Some aim at improving the use of cars with internal combustion engine (e.g., information campaigns on fuel-efficient driving), others aim at disincentivizing the use of private cars (e.g., through comprehensive short-term parking fees). We coded both policy types as targeting private cars, but pursuing different demand-side policy options (improve or shift, see below). While policies that aim at promoting transition away from the use of private cars can also be considered to indirectly target shifts towards other transport modes (e.g., public transport, cycling), we only coded these policies as targeting shifts away from the use of private cars.

  • Regarding instrument choice, policies were coded as economic when they sought to affect household investment or behavioral choices through financial incentives or disincentives, even if the policy itself introduced regulatory changes to achieve this reconfiguration of incentive structures. In other words, the key criterion for the coding decision was the mechanisms by which policies aimed at affecting household GHG emissions, rather than the policy instrument type as such. For instance, the introduction and successive expansion of short-term parking zones and fees in Vienna was correspondingly coded as “economic” rather than “regulatory”, because it increased the costs of using private cars for trips within or to the city.

  • In some cases, striking a balance between ensuring a high resolution and level of detail of the policy mix analysis on the one hand, and maintaining a focus on the overarching picture, necessitated discretionary decisions. For instance, the federally funded program “Climate and Energy Model Regions” leaves municipalities some space to decide what measures they pursue to mitigate climate change and lower dependence on fossil energy. Here, we did not consider every single measure implemented in a municipality, but the main foci of the program, identified through tender specification and measures most frequently deployed across municipalities.

  • The avoid-shift-improve framework is well established in the literature, but classifying policies accordingly required developing some further coding rules (see also Table 1 in the main article):

  • For transport, we coded policies as avoid if they primarily aimed at avoiding or reducing mobility and related emissions, e.g., through promotion of teleworking or compact city structures.

  • Policies were coded as shift if they aimed at promoting transitions from private car use to other transport modes (public transport, cycling, walking) or car-sharing.

  • Policies coded as improve aim at reducing energy consumption and related GHG emissions by fostering choices for smaller or more efficient combustion engines or using sustainable fuels. Following established classification10, we also coded policies encouraging end-user adoption of battery electric vehicles as improve measures, as switching to battery electric vehicles does not constitute a transition to another mode of transport.

  • For housing, measures were coded as avoid if they promote building designs such as passive houses that avoid or drastically reduce demand for space-conditioning services, as well as policies which aim at avoiding new construction altogether, or which seek to avoid demand for space conditioning, lightning and carbon-intensive building materials through adapting dwelling size to household size or shared spaces. We also classified policies as avoid that aim at avoiding household energy demand through behavioral energy saving10.

  • Shift measures encompass policies that enhance the access to or support the switch to renewable energy8, including the electrification of cooking and water heating with renewable energy.

  • Improve measures are policies which aim at improving energy efficiency of new buildings and existing building stock, e.g., via improved building standards, or of household appliances.

Our focus on demand-side policies targeting household GHG emissions implies that we excluded policies focusing on the supply-side (e.g., increasing the share of renewable energy in energy supply) or targeting final business rather than household demand (e.g., incentivizing adoption of battery electric vehicles in corporate fleets), except when policies fell into a gray area. We also excluded policies that may have had major impacts on household GHG emissions without, however, intentionally targeting them as a primary or secondary policy objective (e.g., electricity-price subsidies to cushion energy price hikes for households in 2022). This analytical decision is primarily justified by our interest to better understand which demand-side policy options have been considered and implemented by policy-makers with the aim of curbing GHG emissions. In addition, considering every single policy with a potentially relevant impact on household GHG emissions would have resulted in an unmanageably large policy dataset.

Importantly, we considered major climate policy strategy documents such as Lower Austria’s electric mobility strategy or the overarching federal energy and climate strategy (Mission 2030), even though they often contain declarations of intent which have not consistently been implemented, or where implementation was not consistently traceable. For instance, declarations of intent to increase the share of battery electric vehicles by improving relevant infrastructures may orient policy makers’ infrastructure investment decisions at local policy levels, but investigating such effects was beyond the scope of the analysis. We therefore included policy measures announced in plans or strategy documents, but consistently coded them as “strategy/plan” if we did not find specific measures to implement them, or if the implementation was outside of the scope of the policy analysis (e.g., tracing the realization of individual infrastructure projects such as bicycle lane construction on the district level in Vienna). While this approach runs the risk of overestimating policy action, it allowed us to pinpoint existing demand-side policy mixes and remaining policy gaps more reliably.

A key limitation of this method is that coding and counting policies necessarily compares “apples with oranges”, i.e., it cannot account for the fact that policies vary in scope, effect and importance. Therefore, we complemented our policy mapping with an expert survey among peer researchers to differentiate important policy measures from less powerful ones, which informed the discussion of our results as well as compilation of Figs. 4 and 5. While assessing the effectiveness of policies in terms of achieved GHG emission reductions is notoriously difficult38 and beyond the scope of this paper, our findings may also be useful to explain household GHG footprint emission trajectories as part of future research.

Method 2: GHG footprint calculations for Figs. 1, 2, 4 and 5

To calculate the Austrian consumption-based GHG emissions (footprint) we used the input-output tables of EXIOBASEv3.8.276. We performed standardized input-output modeling to achieve GHG footprints77,78:

$${footprint}=f{({{{\boldsymbol{I}}}}-{{{\boldsymbol{A}}}})}^{-1}{{{\boldsymbol{Y}}}}$$

where \({{{\boldsymbol{Y}}}}\) is the matrix of final demand (from households, government, capital investments, non-profit organizations serving households, and changes in inventories), \(f\) represents the satellite accounts (where we replaced the GHG accounts with official Austrian data), \({{{\boldsymbol{A}}}}\) the technology matrix, and \({{{\boldsymbol{I}}}}\) an identity matrix which has ones on the main diagonal and zeros elsewhere. This modeling approach is described in more detail in and has been used by a number of previous studies79,80,81.

EXIOBASE version 3.8.2 uses nowcasting from 2011 onwards for its underlying monetary data structure, implying that more recent short-term disruptions, like events related to the COVID-19 pandemic, are not reflected in the monetary production, trade, or demand structures. However, GHG extensions are updated up until 2019 for CO2 (based on the Edgar Database) and up until the year 2017 for the other GHGs (based on the PRIMAST database). Moreover, for this study, we have replaced all production-based GHG emissions for Austria with official data based on Statistics Austria and the Environment Agency Austria, using the residence accounting principle82. This includes the direct household emissions from mobility and heating which were taken from separate microcensus surveys of the Statistics Austria collecting data on motor and heating fuels used by private households. EXIOBASE tables provide 200 single products whose GHG footprints of household final demand were aggregated for five larger product groups which are displayed in Fig. 1: mobility, housing and heating, food and beverages, services, and goods, as well as the direct emissions from mobility and heating/housing.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.