From Split Incentives to Shared Responsibility: Evaluation of Germany’s Carbon Cost Allocation Law for Rental Buildings

Introduction

Carbon pricing is a widely used instrument for climate change mitigation and has proven effective to reduce carbon emissions. It is seen as indispensable for reaching the goals of the Paris climate agreement by internalizing externalities and spurring low-carbon tech and investment (see Hepburn, Stern & Stiglitz 2020; Löschel, Sturm & Vogt 2021; IPCC 2022). For example, a meta-analysis of carbon pricing schemes finds that 17 of 21 schemes have delivered immediate emission cuts of 5–21%, despite the low level of prices in most instances (Döbbeling-Hildebrandt et al. 2024). Similarly, Sen et al. (2024) estimate carbon pricing delivers 8–12% average per capita CO₂ reductions across 43 countries, reaching 19–23% after 10 years via fuel switching independent of renewables policies. Looking at the EU Emissions Trading System (EU-ETS 1), firm-level evidence using large, matched samples of regulated and matched non-regulated manufacturing firms show it cut CO₂ emissions by 14–16% without harming economic activity (Colmer et al. 2025). This development was driven by energy-saving investments that lowered emissions intensity rather than shifting production abroad. Moreover, EU-ETS 1 emissions from power and industry installations are around 50% below 2005 levels and on track to achieve the 2030 target of –62% emissions (European Commission 2025). In Germany, the national Emissions Trading System (nETS) and the EU-ETS 1 combined have generated record revenues exceeding €21 billion in 2025, funding climate subsidies and spurring fuel efficiency (Umweltbundesamt 2026).

However, distributional effects pose a key challenge for social fairness and acceptance of carbon pricing, as these policies often exhibit regressive impacts that disproportionately burden low-income households through higher fossil fuel costs. For instance, in France, Berry (2019) quantifies that a carbon tax of €30.5/tCO₂ had a 2.7 times greater impact on the poorest 10% than on the wealthiest 10%, and 1.5 times greater than on the average household. Even within the same income group, carbon pricing effects vary substantially due to diverse building stocks. CO₂ costs particularly hit households with structurally high fossil energy consumption or lower incomes hardest, according to Burger et al. (2022) and Kalkuhl et al. (2024). These fairness concerns contribute to carbon pricing, often lacking majority public support and polls typically show only 30–40% approval, though support rises 20–25% with revenue recycling to low-income rebates (Maestre-Andrés et al. 2019). A German study of 15,000 households by Kaestner et al. (2025) found energy efficiency-based cost sharing most preferred (51% support) over uniform pricing (44%), with preferences unaffected by revenue use or carbon price levels (€25–130/tCO₂).

In the buildings sector, the landlord-tenant problem presents a significant barrier to energy efficiency and carbon pricing effectiveness (Kühn et al. 2024; Kaestner et al. 2025). Landlords’ incentives to minimize upfront maintenance costs often conflict with tenants’ interest in retrofits that lower energy bills (Bird & Hernandez 2012; Gillingham et al. 2012). This principal-agent dilemma arises because landlords control investment decisions and tenants typically bear the heating costs, leading to split incentives and underinvestment in efficiency measures (see Bird & Hernandez 2012; Gillingham et al. 2012; Ástmarsson, Jensen & Maslesa 2013; Horng & Novikova 2018). These barriers are particularly severe in low-income rental housing as landlords capture few of the benefits from energy efficiency investments while tenants lack authority over building upgrades (Bird & Hernández 2012). As a result, landlords may delay renovations or pass costs to tenants via higher rents, exacerbating housing cost pressure (Ástmarsson, Jensen & Maslesa 2013; Behr et al. 2024). Germany faces acute challenges here, with the highest share of renters in the EU (53%) (Eurostat 2024a), comparably high heating emissions from buildings and heating and a buildings sector that continues to miss its climate targets (Umweltbundesamt 2024a; Expertenrat für Klimafragen 2025). Consequently, energy retrofit rates in German residential buildings have stagnated below 1% for over a decade despite substantial subsidy programs for energy-efficient retrofits (Galvin 2024; Umweltbundesamt 2019). Notably, when landlords cover heating costs, they invest more in efficiency (Melvin 2018).

International experience suggests that split incentives are addressed in many countries through a combination of regulatory, financial, and tenancy-law instruments. Castellazzi, Bertoldi and Economidou (2017) give an overview across approaches in European countries. Regulatory approaches address split incentives by shifting responsibility for energy performance to property owners through minimum energy performance standards, while reforms to rent and condominium laws can enable cost-sharing arrangements, green leases, and decision-making processes that facilitate energy renovations. Information-based instruments, such as energy performance certificates and individual metering, aim to reduce information asymmetries, improve awareness of energy consumption and costs, and provide incentives for both landlords and tenants to support efficiency improvements. Financial solutions seek to overcome upfront investment barriers through subsidies, tax incentives, and innovative financing mechanisms such as on-bill financing and Property Assessed Clean Energy (PACE), which can align investment incentives by linking repayment to energy savings or the property rather than the individual owner. Voluntary mechanisms, particularly green leases, attempt to overcome split incentives by allowing landlords and tenants to share the costs and benefits of energy efficiency investments, although their adoption remains limited in many European contexts. Beyond these instruments, countries have also experimented with different tenancy-law arrangements to address the allocation of energy costs. Sweden’s inclusive rent system shifts heating costs to landlords, strengthening incentives for energy-efficiency retrofits, although it may weaken tenants’ incentives to conserve energy (Reutter 2025).

Germany’s Carbon Dioxide Cost Allocation Act (CO₂KostAufG 2022) represents a novel regulatory response to the landlord–tenant dilemma by explicitly reallocating carbon price burdens for space heating between landlords and tenants. Introduced in 2023, the law links the distribution of heating-related CO₂ costs between landlords and tenants to building-level emissions performance, with the aim of strengthening renovation incentives while limiting regressive impacts on tenants. As such, the CO₂KostAufG constitutes a rare example of a rule-based, ex-post incidence adjustment embedded within an emissions trading framework, and is of particular relevance for the broader EU debate on socially balanced carbon pricing and the implementation of EU-ETS 2 in buildings. This paper presents a process evaluation of the German CO₂ cost-splitting scheme. Given that the scheme was only introduced recently, it is too early to assess its long-term impacts, such as investments in building energy efficiency or renewable heating systems. Instead, the paper examines the implementation of the scheme, focusing on whether the cost-splitting provisions are applied as intended, the challenges encountered by market participants, and the resulting distribution of CO₂ costs between landlords and tenants.

The German Carbon Dioxide Cost Allocation Act

The German Carbon Dioxide Cost Allocation Act took effect in January 2023 and was first applied in 2024 to the previous heating period 2023 (CO₂KostAufG 2022). The law operates alongside Germany’s nETS and the EU-ETS 1, which jointly price fossil heating fuels such as natural gas (nETS), heating oil (nETS), and district heating (EU-ETS 1, nETS for smaller installations).

The policy aims to promote a fairer distribution of CO₂ costs in rental buildings while addressing the landlord–tenant split incentive. Its objectives are to encourage energy-efficient renovations and the adoption of low-emission heating systems by landlords, preserve incentives for energy-efficient behavior by tenants, and ultimately reduce heating-related emissions in the building stock. To this end, the law introduces a cost-sharing mechanism that links the allocation of carbon costs to the realized emissions performance of individual buildings.

Under the CO₂KostAufG, CO₂ costs are distributed using a 10-tier model based on annual CO₂ emissions per square meter of living space. Emissions are calculated using reported fuel consumption, building living space, and fuel-specific emission factors. Buildings with higher emissions intensity are assigned classes, in which landlords bear a larger share of CO₂ costs, whereas lower-emission buildings shift a greater share of costs to tenants (Figure 1). At the extremes of the model, tenants bear the full CO₂ costs in the best-performing class I, while landlords cover up to 95% of the costs in the worst-performing class X.

Figure 1

The CO₂KostAufG contains several exemptions that limit or exclude landlords’ cost-sharing obligations. Under § 9(1), the landlord’s share of CO₂ costs is halved when legal or regulatory constraints restrict energy efficiency improvements, such as in listed buildings, areas subject to preservation statutes, or where municipal regulations mandate district heating and limit landlords’ control over the heating system. In addition, the law does not apply to CO₂ costs from fuels invoiced before 2023, irrespective of the actual consumption period (§ 11(2)). This applies mostly to heating oil purchases before 2023. Finally, buildings newly connected to district heating after the onset of 2023, are fully exempt from the cost allocation mechanism (§ 2(3)).

Methodology

This analysis used a mixed-methods approach which is visualized in Figure 2. It consists of four methodological steps: the characterization of the German rental housing stock, the calculation of the modelled and observed CO₂ cost allocation between landlords and tenants, and the analysis of efficiency of the approach and implementation challenges.

Figure 2

As the CO₂ Cost Allocation Act entered into force only in 2023, the evaluation focuses on the early implementation of the legislation. We combine quantitative and qualitative methods to assess three complementary dimensions. First, the operational functioning of the allocation mechanism is analysed by examining whether the legislation is being implemented within existing heating cost billing procedures and identifying practical implementation challenges. Second, the distributional effects of the legislation are assessed by comparing the expected allocation of CO₂ costs with the observed allocation derived from billing service providers and stakeholder information. Third, stakeholder interviews and the online survey are used to evaluate the practical applicability of the legislation and to identify implementation barriers perceived by landlords, property managers and billing service providers.

Results and Discussion

Characterization of German rented housing stock

Figure 3 provides an overview of the distribution of rented units by building type, main energy carrier used for heating and ownership type. The largest group of tenancies is found in multi-family homes (MFHs) (83%), followed by two-family homes (TFHs) (10%) and single-family homes (SFHs) (8%). Regarding heating types considered under the CO₂KostAufG, the dominating energy carrier is natural gas heating (60%), as central (47%) or individual heating units (13%), especially in MFHs. District heating makes up 22%, mostly in MFHs. Heating oil makes up for 13% of the overall rented units in Germany, with the highest share in TFHs (30%) and SFHs (23%). Moreover, private individuals are by far the largest landlord group in all ownership types.

Figure 3

CO₂ cost distribution

Modelled distribution

Figure 4 shows the results of the modelled distribution of CO₂ costs using energy cost certificates. The results show that across all tenancies, landlords are modelled to account for approximately 52% of the costs for multi-family dwellings, while tenants cover approximately 48% of the costs.

Figure 4

Calculated distribution based on data from billing service providers

Figure 5 shows the distribution of buildings by emission class based on the data of three different billing service providers. The figure shows that a large proportion of buildings fall into the first classes, particularly for natural gas and district heating. Overall, the evaluation results in CO₂ cost shares for landlords of 26%, 27%, and 26%.

Figure 5

Distribution based on the survey

In the survey, participating companies provided information on the estimated cost distribution in their managed building portfolios (Figure 6). The most frequently selected response category was a landlord share of 20–30%. Over 83% stated that their portfolio falls into one of the first four emission classes of the CO₂KostAufG. Less than 5% reported landlord shares of over 60%. This results in a total landlord share of 33%.

Figure 6

Comparative overview

A comparison of the results of the shares calculated in the previous sections shows that the modelled distribution based on energy performance certificates yields a significantly higher share of CO₂ costs borne by landlords than the other two data sources used (Table 3). This suggests that, on average, energy performance certificates record higher energy consumption than the consumption data from 2023.

Table 3

Overview on the CO₂ cost distribution between landlords and tenants.

	ENERGY CERTIFICATES	BILLING SERVICE PROVIDERS	SURVEY
Share Landlords	52% (only MFH), 60% (MFH, SFH, TFH)	26%, 27%, 36% (only MFH)	33%
Share Tenants	48% (only MFH), 40% (MFH, SFH, TFH)	74%, 73%, 64% (only MFH)	67%

Table 3 reveals substantial discrepancies between the modelled and observed distribution of CO₂ costs between landlords and tenants. The following section analyses the underlying reasons for these differences and distinguishes between temporary and systematic effects.

Methodological implications of weather normalization for policy calibration

The discrepancy between the modelled and observed distribution of CO₂ costs shown in Table 3 can be explained by two fundamentally different mechanisms. The first is temporary and results from unusually low heating demand during the 2022–2023 energy crisis (see Roth & Schmidt 2023; Jamissen et al. 2024), although Singhal et al. (2026) found a large intention-to-action gap in reducing heating consumption. This effect influences the annual distribution of buildings across the emission classes of the CO₂KostAufG but is not expected to persist over longer time periods. They therefore do not fundamentally affect the validity of the policy calibration.

The second mechanism is systematic and results from the methodology used to weather-normalize energy consumption in German consumption-based energy performance certificates (EPCs). Consumption-based EPCs are designed to compare the intrinsic energy performance of buildings independently of the weather conditions during the observation period. To achieve this, measured annual heating energy consumption is adjusted using climate correction factors (Klimafaktoren) published by the German Meteorological Service (DWD).

This approach is appropriate provided that the reference climate remains representative of the climatic conditions under which the corrected energy consumption is interpreted. Under this assumption, weather-normalized consumption provides an approximately unbiased estimate of the long-term operational energy consumption of the building stock. However, this assumption is increasingly violated as a consequence of climate change. Germany has experienced a sustained increase in winter temperatures over recent decades, resulting in substantially lower heating demand than during the reference period used to derive the DWD climate correction factors, which, in turn, were used to define the ten levels of CO₂-cost sharing within the CO₂KostAufG. Lower real energy consumption as a consequence of climatic warming therefore leads to buildings systematically entering CO₂-cost-sharing levels where the tenants have to pay a higher share of the CO₂-costs.

Importantly, this finding does not imply a weakness of the operational design of the CO₂KostAufG. The legislation intentionally allocates CO₂ costs on the basis of realised annual fuel consumption obtained from the existing heating cost billing process. This considerably simplifies implementation, avoids additional weather normalization during billing, and ensures that the allocated costs correspond to the CO₂ emissions that actually occurred during the billing period. Over multiple years, warmer and colder winters naturally average out in the operational data. The systematic bias therefore arises exclusively in the ex-ante calibration of the expected cost distribution and not in the practical implementation of the legislation.

Application to other EU Countries

For an EU-wide perspective, the EU-SILC 2023 microdata provides information linking tenancy status with heating type used in the dwelling. This allows a preliminary assessment of which countries might be more amenable to implementing a similar system. Figure 7 presents the share of tenant households that live in dwellings with central and district heating across EU member states.

Figure 7

From this analysis, it becomes evident that countries with high proportions of tenancies with central heating systems and district heating would likely find it easier to implement a CO₂ cost allocation system similar to Germany. Conversely, countries with predominantly individual storey heating, such as Spain, Portugal, and Malta, may face higher implementation challenges, as heating costs are directly borne by individual households and there is a limited existing infrastructure for distributing CO₂ costs collectively.

This result suggests that while the German model is promising, its direct transferability depends on housing structure, tenancy rates, and existing heating infrastructure. Further research is needed to quantify the potential applicability and identify country-specific barriers.

Conclusions

Our analysis shows that the German approach to splitting carbon costs in rented residential buildings is applied across most segments of the rental market. A key factor behind its successful implementation is that the scheme is integrated into existing sub-metering and billing procedures, minimizing additional administrative effort. However, this integration also means that the cost split is based on actual energy consumption rather than standardized consumption levels. As a result, landlords may bear a larger share of the CO₂ costs when tenants consume relatively little heating energy—for example, because they underheat their homes due to energy poverty.

However, the mechanism performs less effectively in certain market segments, particularly in single-family homes and buildings with individual gas heating systems in each flat. In these cases, the administrative procedures and allocation rules have shown to be less suitable, indicating that the design of the mechanism does not fully accommodate all building and heating system configurations.

A key finding of the evaluation concerns the equity of the programme: The distribution of carbon costs between tenants and landlords differs substantially from the original policy objective. The legislation aimed to achieve an approximately equal split of total carbon costs across the rental sector. In practice, however, tenants currently bear around 70% of the total carbon cost burden.

Two main factors explain this discrepancy. First, the sharp increase in energy prices led many households to reduce heating demand. As a result, a larger share of buildings now falls into lower-emission categories within the cost-sharing scheme, where tenants are responsible for a higher proportion of the carbon costs. Second, technical limitations in the weather correction procedures used in German energy performance certificates affect the modelled cost distribution and therefore distort the allocation of carbon cost shares.

Overall, the German model provides an interesting policy approach for other countries seeking to introduce carbon cost sharing between landlords and tenants. However, successful transfer requires careful consideration of national billing practices, building stock characteristics, and the design of building energy certification systems.