Deep energy efficiency renovation of Germany’s residential buildings: is this as economically viable as Germany’s policymakers and popular promoters often claim?

Abstract

Germany needs to steeply increase the rate of deep energy-efficiency renovation of its old buildings, to meet climate goals. The government has long maintained that deep renovation to a specified “minimum” energy efficiency standard is economically viable: that the costs are paid back, through energy savings, over the technical lifetime of the energy-efficiency measures. Many private and public organisations support and promulgate this view. It accords with a stream of academic literature which suggests that the under-adoption of energy efficiency measures is a paradox indicating market failures and economically irrational behaviour by property owners. This paper offers cost–benefit analyses of 44 case study scenarios to test whether deep renovation in Germany pays back in monetary terms. These include both specific buildings and Germany-wide averages of classes of buildings. It uses current construction, finance and energy costs, and takes account of inflation, discount rates and opportunity costs. None of the scenarios are economically viable in monetary terms, and the average payback after 25 years is around 22.5%. Sensitivity analyses suggest payback would only be achieved using improbable parameter values. Energy-efficiency renovation is necessary but promoting it needs to take account of these realities.

Introduction

Buildings are responsible for around 15% of Germany’s CO2 emissions, most of which are from space heating (Bundesregierung, 2019). Because very high energy efficiency standards are demanded for new builds, CO2 emissions from the buildings sector reduced by 44% between 1990 and 2023. However, to meet Germany’s climate goals, emissions from buildings will need to fall by a further 23% by 2030, and to nearly zero by 2050 (Bundesregierung, 2022a, 2022b). Most of this will have to come from deep energy efficiency upgrades of the country’s large stock of older buildings.

The German government has long maintained that it is economically viable to renovate old buildings to a specified “minimum” energy efficiency standard. This standard is equivalent to achieving energy consumption no greater than 40% higher (worse) than the new-build standard for a building of the same dimensions. Typically, this puts the minimum standard for comprehensive renovations at about 70 kWh/m²/y. This invariably involves applying thick wall insulation, installing double- or triple-glazed windows, and insulating the basement ceiling and the roof. I call this “deep” renovation, or “deep energy efficiency” renovation in this paper to distinguish it from minor energy efficiency improvements such as draft proofing, adjusting boilers and setting appropriate daily heating periods. Deep renovation may also include the transition to heat pumps (Eyre et al., 2023), which can substantially reduce energy demand because of their high coefficients of performance. However, heat pumps generally perform well only if the building envelope is already at a standard of about 70 kWh/m²/y or better (Terry & Galvin, 2023). Hence they cannot be used as a substitute for deep energy-efficiency renovation.

The government has long argued that for renovations to this standard the financial savings from reduced energy consumption will compensate for the energy-efficiency improvement costs over the technical lifetime of the energy-efficiency upgrade measures (Galvin, 2010; Galvin & Sunikka-Blank, 2017). For standards higher than this, which cost even more to achieve, the government offers a range of subsidies to help cover the extra costs.

The claim that deep renovation is economically viable resonates with a narrative in economics literature, namely that energy efficiency technology pays for itself through the benefits it brings (e.g., Jaffe & Stavins, 1994a, 1994b; Solà et al., 2021). It also resonates with much popular promotional material on the economic benefits of increasing the energy efficiency of Germany’s older buildings (e.g., Deutsche Umwelthilfe, 2017; DENEFF, 2023; Dämmen-lohnt-sich, 2023; co2online, 2023).

One might expect, then, that building owners throughout Germany would be rushing to do deep renovation of their buildings. However, the annual rate of energy-efficiency renovation in Germany has been disappointingly low. The German Energy Agency estimates it at below 1% per year, meaning that only 1% of highly energy inefficient buildings are getting deep renovation each year, and notes that it must increase “substantially” (erheblich) to play its part in meeting Germany’s climate goals (dena, 2022).

A number of reasons have been suggested for the slow uptake of deep energy-efficiency renovation in Germany’s older buildings. These include lack of upfront finance (Bertoldi et al., 2021); the landlord-tenant dilemma, often called the principal agent problem (März et al., 2022); hesitancy due to the promise of higher future energy prices (Kumbaroğlu & Madlener, 2012); and a cluster of psychological, social and administrative barriers (Baumhof et al., 2018; Klöckner & Nayum, 2017). This paper investigates a further possible reason: that it might not be economically viable after all. Since it is widely proclaimed in Germany that property owners get their money back, through energy savings, over the technical lifetime of the energy-efficiency upgrade measures, we need to examine the empirical evidence of whether this is the case and if so, to what extent. This paper therefore reports the results of 44 case study scenarios of the likely costs and expected benefits, through energy and CO2 tax savings, of energy-efficiency renovation in Germany. 29 of these case studies are specific buildings, but 15 are Germany-wide averages of the most common classes of buildings needing deep renovation.

There are at least four different types of benefit which an energy efficiency upgrade can bring: direct monetary savings from reductions in energy and CO2 tax bills; indirect benefits to the occupants in terms of increased comfort and wellbeing; the indirect benefit to the property owner of increased value for their building; and wider societal benefits such as reduced CO2 and particle emissions. While all of these are important, this paper focuses on the first, for two reasons. First, whether or not a deep renovation is likely to pay back directly affects the budgets of property owners, and therefore plays a large role in whether or not they choose to renovate.

Second, it is the basis of the government’s claim that deep renovation to the basic minimum standard is economically viable. As noted below, Germany’s federal building regulations can only mandate energy-efficiency standards that can be achieved if the monetary benefits equal or surpass the costs. The government also uses this as a lever to attempt to persuade property owners to renovate. It is therefore essential to check whether energy-efficiency renovations do, in fact, achieve this balance of costs and benefits.

The paper also takes this a step further in that it moves beyond a simple summation of costs and projected benefits, to employ cost–benefit analyses based on net-present-values (NPVs) and including realistic finance costs (loan interest rates), investor discount rates, energy price inflation rates and opportunity costs.¹ It also provides sensitivity analyses to show the effects of varying these. The study also provides rule-of-thumb estimates in straightforward terms for property owners who prefer to avoid the complexities of NPV estimates.

The remainder of the paper is as follows. Section "Literature" briefly reviews literature relevant to the German government’s discourse on economic viability, then similar discourse in grey literature, then academic literature on the economic viability of energy efficiency upgrades. Section "Method" explains the method for the empirical analysis. Section "Results" gives the results of the analysis, Section "Discussion" discusses the implications of these results, and Section "Conclusions" concludes.

Literature

The idea that deep energy-efficiency renovation of Germany’s (and other countries’) buildings always pays back is evident in three domains: policy discourse; grey literature; and some academic literature. This review mentions only a small, representative sample of the very large literature on each of these.

Policy discourse

Successive German governments have long maintained that it is always economically viable to do deep energy efficiency renovation on buildings that were constructed in the years and decades before the late 1970s, when energy efficiency standards became mandatory for new builds (Galvin & Sunikka-Blank, 2017; Galvin, 2010, 2012, 2014). More precisely, the government has periodically contracted research institutes to calculate the minimum standard of energy efficiency to which it was economically viable to upgrade. This standard has depended to some extent on the geometric features of the building,² and on construction costs, finance costs and energy prices. It averaged around 100 kWh/m²/y until the mid-2000s and is currently around 70 kWh/m²/y. The contractors calculated that the costs of renovating an old building to that standard would be paid back, through energy savings, within the technical lifetime of the renovations, generally taken to be 25 years. The government mandated that all comprehensive renovations must reach that standard or better. “Comprehensive” renovation is a legal definition in Germany for a situation where all the major features of a building envelope are being renovated, i.e. the walls, windows, roof and basement ceiling.”

This minimum standard of energy efficiency is a legally binding restraint on what the government could mandate in the building code. A law known as the Energieinsparungsgesetz (Energy saving law – EnEG), passed in 1976, stated that the government could only mandate energy efficiency upgrades that were “wirtschaftlich vertretbar” (economically justifiable or viable). The energy efficiency building code, known as the Energieeinsparverordnung (Energy saving ordinance – EnEV) therefore mandated the energy efficiency standard that the contractors had calculated to be economically viable. More recently this has been embodied in a new building code law, the Gebäudeenergiegesetz (GBG – Building Energy Law).

The government has always accepted that renovating to a higher energy standard than this (i.e., a lower number of kWh/m²/y) was not likely to be economically viable. It therefore runs a subsidy scheme, via the German Development Bank (KfW – Kreditanstalt für Wiederaufbau), for renovations that achieve exceptionally high standards, currently around 27 kWh/m²/y.

There have always been, however, two qualifications, or caveats, on this system. First, the contractors are not allowed to include the so-called “anyway costs” (called in German variously Sowiesokosten, Umsonstkosten or Ohnehinkosten) in the calculations. It is assumed that these costs would have to be paid “anyway”, to bring the building to a modern, robust physical standard, so they should not be counted as energy efficiency costs. An example of a government contractor’s definition of “anyway” costs is given in Appendix 1 (author’s translation from Hinz and Enseling, 2021, p. 26).

This is controversial because it brackets many costs out of consideration which would certainly have to be paid in real life. For example, half the costs of the scaffolding, all architects’ fees, half the costs of door replacement and roof upgrading, and most of the window replacement costs are excluded from the calculations regardless of how good or bad these features were before the energy-efficiency upgrade. Nevertheless, in the empirical section below I adopt this approach, simply to investigate whether energy-efficiency upgrades are economically viable given this caveat.

The second caveat is that the calculations ignore the prebound effect, a widely occurring phenomenon in residential buildings in Germany and other European countries, first identified by Sunikka-Blank and Galvin (2012) and subsequently extensively studied in Germany and other countries (examples are: Desvallées, 2022; Gróf et al., 2022; Karpinska & Smiech, 2020; Weber & Wolff, 2018; Geraldi & Ghisi, 2020; Guerra-Santin et al., 2016). The prebound effect is the phenomenon that the average actual heating energy consumption in older, energy-inefficient dwellings is about 30–40% lower than the official energy rating which is based on engineering calculations. There may be many reasons for this, such as wrong technical assumptions in the engineering calculations (Cozza et al., 2021), users under-heating to save money (Berger & Höltl, 2019; Tel et al., 2016), household practices not being taken into account in methods of calculating energy efficiency ratings (Gram-Hanssen, 2014; Steemers & Yun, 2010), or various combinations of these (Galvin, 2023a, 2023b).

Whatever the reasons for the prebound effect, the important point for this discussion is that ignoring it compromises the credibility of the findings of the government’s contractors. By ignoring it, they inadvertently exaggerate energy savings by around 50–80%. For example, a typical apartment built in the 1960s has an official energy rating of around 145 kWh/m²/y. With a prebound effect of just 30% the actual pre-renovation consumption is around 102 kWh/m²/y. If the apartment is renovated to the minimum standard of 70 kWh/m²/y, the occupants will therefore not reduce their energy consumption by 75 kWh/m²/y but only by 32 kWh/m²/y, which is only 43% of the assumed savings. The renovation will therefore not pay back in 25 years. The investor will get only 43% of the money back that they expected to get.

Unfortunately, this mathematical glitch has severely compromised the credibility of the German government’s claim of economic viability over the past two decades. In the empirical section below I include prebound effects in most but not all of the scenarios, and for others I calculate both with and without prebound effects.

Moreover, over the past three years construction costs have increased steeply, by about 43% since 2020 and 34% since 2021 (Statista, 2023), while interest rates have increased from around 1% to over 4%, thereby increasing finance costs by over 300%. Meanwhile the cost of natural gas, which is the dominant heating fuel in Germany, increased but then settled at about 50% above its pre-Ukraine-war level, due to strong political pressure to keep it affordable (Bundesregierung, 2022a). Therefore the increase in energy prices has not been sufficient to offset the increases in construction and finance costs. Hence, even the Institute for Housing and Environment, one of the government’s regular contractors, began to maintain, in 2022, that energy-efficiency renovation is no longer economically viable – even if “anyway” costs and prebound effects are ignored (Hinz & Enseling, 2022).

Grey literature

A number of Germany’s online publications and websites of firms and organisations maintain that deep energy-efficiency upgrades are economically viable, almost always emphasising that “anyway” costs should not be counted in the calculations, a position which this paper accepts, but also usually ignoring prebound effects. Just four examples are given here.

The NGO Deutsche Umwelthilfe (German Environmental Aid) produced a booklet “Questions and Answers on the Economic Viability of Energy-Efficiency Renovation” in 2017 (Deutsche Umwelthilfe, 2017). The section dealing with cost accounting very carefully explained which costs to include or exclude from the calculations. It also advocated that pre-renovation consumption be estimated as the official energy rating (rather than the actual consumption of the occupants, i.e., ignoring prebound effects). Another organisation, Dämmen-lohnt-sich (Insulation Pays Back) cited a report from 2014 by FIW, a Munich research institute funded by the Association of Insulation Industries (Gesamtverband Dämmstoffindustrie). It listed the payback times of individual insulation measures (walls, roof, cellar ceiling and attic floor). It stated that payback times for these range from a worst case of 32 years to a best case of 8, with an average of 15 (Dämmen-lohnt-sich, 2023). Again, the assumptions were that “anyway” costs are excluded from the calculations and pre-insulation consumption is identical with the building’s official energy rating (again ignoring prebound effects.

Another example is the NGO co2online. Using the same assumptions, co2online claimed to have averaged the costs and benefits of a large number of renovations reported by its members in the pre-covid years and maintained that these paid back (co2online, 2023).³

As noted above, construction costs have increased steeply since 2020, but the idea that energy efficiency renovation always pays back continues to be promoted. This is seen in a recent publication by the Deutsche Unternehmensinitiative Energieeffizienz⁴ (DENEFF—German Industries Initiative Energy Efficiency). In its 2023 publication Finanzinnovationen für die Wärmewende (Finance innovations for warmer homes), DENEFF (2023) states that “As a general rule, even ambitious energy-efficiency upgrades are economically viable over the lifetime of the building⁵”. This claim is supported only by a footnote referring to a study by the Institute for Housing and Environment, namely Hinz and Enseling (2022), already mentioned above. This study gives the costs and energy savings from three different levels of energy-efficiency renovation for each of three different standard house types found throughout Germany. No actual cost–benefit analyses are performed, but the authors conclude: “Without subsidies, ambitious and climate-target-compatible modernisations are not yet economically viable on an area-weighted average under the assumed framework conditions, though a number of exceptions exist.⁶” (emphasis added by author).

This hardly sounds like an endorsement of DENEFF’s claims. In Section "Results" I give the results of my own cost–benefit analyses of the nine case study scenarios from this study (alongside some 30 others), updating the energy, construction and finance costs to 2023 levels.

There is, then, a major stream of grey literature that continues to promote the notion that energy-efficiency renovation is economically viable, though it does not support this with hard evidence, and the evidence it refers to sometimes even shows the opposite.

Academic literature

There is a sizeable stream of academic literature that assumes energy efficiency pays back. Many if not most of these papers ground this assumption in their interpretation of early works by Jaffe and Stavins (1994a, 1994b), though the ideas in these were largely prefigured by Fisher and Rothkopf (1989). Examples dealing with energy efficiency upgrades in buildings include Adam et al. (2005), Gerarden et al. (2017), Schleich (2004), Giraudet (2020), Müller and Berker (2013), Solà et al. (2021), Myers (2020) and many others.

Nevertheless, Jaffe and Stavins (1994a, 1994b) are not concerned to provide hard evidence for the economic viability of energy efficiency upgrades in buildings, and therefore do not present case studies or compendia of case studies. Rather, they “take it as given” that energy-efficiency upgrades – which they call “improved thermal insulation materials” – are economically viable. They call the gap between actual and economically viable energy efficiency the “energy efficiency gap” (Jaffe & Stavins, 1994b), and property owners’ failure to invest in economically viable energy efficiency the “energy efficiency paradox” (Jaffe & Stavins, 1994a). The energy efficiency gap and paradox are said to be market failures, due to irrational consumer behaviour and/or non-market barriers.

It is important to see how this plays out in subsequent studies. As an example, Schleich et al. (2019) investigate the effects of behavioural economic factors on adoption of energy efficiency in buildings (and other spheres) in eight European countries. They find evidence that “individuals who are more loss-averse, or more risk-averse, or who exhibit a lower time discount factor” are less likely to have adopted energy efficiency technologies, though they admit the evidence is not fully robust. They assume throughout that there is an “energy efficiency paradox”, adopting this notion from writers such as Allcott and Mullainathan (2010), Allcott (2011), Gerarden et al., (2015, 2017) and Ramos et al. (2015), who in turn adopted the phrase from Jaffe and Stavins (1994a). While their findings indicate that energy-efficiency renovation can be inhibited by behavioural economic factors, this does not necessarily support the idea that it is an economically viable activity that is being under-used. With respect to deep energy efficiency renovation of Germany’s buildings (and of old buildings more widely), there needs to be more empirical research to see if it is economically viable.

Gerarden et al. (2017) find that "energy-efficient technologies offer promise for reducing the costs and environmental damages associated with energy use, but these technologies appear not to be used by consumers and businesses to the degree that would be justified, even on the basis of their private financial net benefits.” They provide a critical review of economic models used to investigate the energy efficiency paradox – which they re-name the “social energy-efficiency gap” – and include technologies that pay back for society as a whole even if they do not pay back for the individual investor. However, they do not offer examples of deep energy-efficiency renovation that pay back in direct monetary terms for the building owner.

On the other hand, Belaïd et al. (2021) analyse a database of 1,400 dwellings collected in 2013, representing a cross-section of dwellings in France. Their analysis focuses on the costs and benefits of individual renovation measures, such as wall insulation, loft insulation, window replacement, etc. They give estimates of the costs and benefits of each of these, showing that some pay back and others do not. The study indicates that with 2013 prices, wall external insulation is the least economic measure. This is an important finding because to meet Germany’s demand for deep renovation external wall insulation is almost always required.

Giraudet et al. (2020) focus on faults and failures in the renovation process, such the use of low-quality materials and post-renovation rebound effects, where households overheat because it is now cheap to do so. Both of these would reduce the energy savings below what was fed into the cost–benefit calculation. We need, however, empirical studies of competently performed deep renovation that is not followed by large rebound effects, to test whether it is economically viable if done properly and used efficiently.

A review by Solà et al. (2021) summarises the range of approaches to the energy-efficiency gap and paradox and classifies its causes as market failures, principal agent problems (Davis, 2011; Gillingham & Palmer, 2014) and “other factors”, including social norms (Liu et al., 2016), procrastination (Lillemo, 2014) and personal experience (Franke et al., 2012). Market failures include hidden and transaction costs, economic myopia (short-sightedness), lower-than-efficient energy prices, and capital market problems. Again, however, there is no reference to a body of work that empirically investigates the direct monetary costs and benefits of specific cases of deep energy-efficiency upgrades.

An important insight highlighted in Solà et al. (2021) is that there are benefits in energy-efficiency renovation that accrue to society rather than the property owner: “decreases in GHG emissions and other local pollutants and the subsequent health benefits”, as explored, for example, in Corradini et al. (2014) and Markandya and Rübbelke (2012). While this is important, it still does not answer the question of whether deep energy efficiency renovation pays back in monetary terms for the property owner. A related question is whether deep renovation is an economically efficient way to reduce greenhouse gas emissions – an issue touched on below, empirically, in Section "Results".

It is also important to think critically about the principal agent problem, also known as the landlord-tenant dilemma, where the landlord pays for the energy-efficiency upgrade but the tenants get the benefits, from reduced energy costs. This is often used to support the notion of the energy-efficiency gap as, it is argued, it makes it uneconomic for a landlord to renovate even when renovation is technically economically viable (Frederiks et al., 2015; Gerarden et al., 2017; Jaffe & Stavins, 1994b; Linares & Labandeira, 2010). This brings us back, therefore, to the question of whether deep energy efficiency renovation is economically viable in the first place.

In summary, the German government has long maintained that deep renovation to the minimum energy efficiency standard is economically viable, and a number of industry and interest groups in Germany also maintain this. Literature on the economic benefits of energy efficiency improvements in general terms tends to be optimistic about their economic viability. However, it is difficult to find literature that has systematically and empirically assessed the economic viability of deep energy-efficiency retrofits. There are good examples of studies that explore the societal value of energy efficiency upgrades in buildings and the problems landlords might have in passing on costs to tenants, but a dearth of studies that look at the direct economic benefits of deep renovation. The next sections seek to fill this gap by offering cost–benefit analyses of a large number of case studies in Germany, including both individual buildings, and country-wide averages.

Method

The case study buildings and scenarios

A number of case studies were gathered from a range of sources, in which work had been done to estimate the costs of energy-efficiency renovation of old residential buildings to a range of standards, in most cases including the legal minimum standard of 70 kWh/m²/y and in some cases higher standards such as 50, 40, 27 and 20 kWh/m²/y. In most of these cases the officially estimated pre-renovation consumption was known, and for some of these prebound effects were estimated based on national averages for similar buildings. In other cases the actual pre-renovation consumption was known.

Table 1 lists the input parameters for the cost–benefit calculations. Table 2 lists the case studies and the input data used in the calculations of economic viability. The cases listed in the table with floor area 1 m² are those where the cost estimates were given on a per m² basis.

Table 1 Input parameters for cost–benefit analyses. Note that all of these can be changed by the user

Table 2 Case study buildings and scenarios. In all cases the energy price for gas is assumed to be 0.115 €/kWh, electricity price 0.358 €/kWh, interest rate on finance 4%, property owner's discount rate 6%, and alternative investment return 7% (for calculating opportunity costs)

Energy prices for 2024 were based on data for January 2024 from 137 gas providers and 127 electricity providers, supplied by Germany’s consumer protection institute (Verbraucherzentral, 2024). From this I use the average gas price of 0.115 €/kWh and average electricity price of 0.358 €/kWh. Other parameter values used are: interest rate on finance 4%, property owner's discount rate 6%, alternative investment return 7% (for calculating opportunity costs⁷), and CO2 tax rate 30 €/tCO2 (as is current in Germany). It is assumed that 10% of the renovation cost is paid up-front and the rest is borrowed. This reduces the effective costs as it keeps opportunity costs to a minimum in a situation where loan finance is still cheaper than the likely return on an alternative investment. However, these parameter values were varied in sensitivity analyses. The target energy consumption is given in column 3 of Table 2.⁸

Going through the list of scenarios in Table 2, scenarios 1–8, the Aachen case studies, were performed by the author. They are based on two typical four-apartment buildings in Aachen. In scenarios 3 and 7 the gas boiler is replaced by a heat pump, and in scenarios 4 and 8 rooftop photovoltaics are then added. An energy performance certificate from 2022 was used to estimate theoretical consumption of building 1, and it was assumed that that for building 2 was similar as they are of the same age and design, though building 2 is about 10% smaller.

Scenarios 9–20 are updates of scenarios for renovating three different apartment buildings to a range of standards, developed by the Institute for Housing and Environment (IWU – Institut Wohnen und Umwelt) in an official report for the city of Augsburg (Enseling et al., 2020). I updated the gas price, construction and finance costs to current levels for the cost–benefit analyses.

Scenarios 21–24 are Germany-wide averages provided by the research institute Arbeitsgemeindschaft für Zeitgemäßes Bauen e.V. (ARGE—Working Group for Contemporary Construction⁹), developed in 2020. The advantage of these scenarios is that they refer to the average sized apartment building in Germany rather than just one building, which might not be as representative. Again, I updated the gas price, construction and finance costs to current levels for the cost–benefit analyses.

Scenarios 25–33 are the case studies from the Institute for Housing and Environment mentioned in Section "Literature", covering three different Germany-wide house types, each renovated to three different standards. I updated these to 2023 prices but retained the investigators’ method of ignoring prebound effects (to check whether the DENEFF reference to this was credible on its own terms).

Scenarios 34–39 are based on Germany-wide averages for buildings containing “small” apartments (average floor area 65.14 m²), provided by a consortium of 6 research institutes and NGOs including the Fraunhofer Institute and the German Energy Agency (Bundesverband neue Energiewirtschaft, 2018). These data are particularly useful because of their country-wide scope and the credibility afforded by the participation of multiple institutes in data gathering and fact-checking. I updated the gas price, construction and finance costs to current levels for the cost–benefit analyses.

Scenarios 40–44 are Germany-wide averages for “large” apartment buildings (called “GMFH”—Großmehrfamilienhäuser), also based on data from Bundesverband neue Energiewirtschaft (2018). For both these datasets I have updated renovation costs to 2023 levels.

I include 3 sensitivity analyses. The first gives the effect on percentage payback time over 25 years, of zero discount rates and zero opportunity costs, for a selection of 6 cases. These correspond to a back-of-the-envelope calculation rather than a NPV calculation.

The second gives percentage payback in 25 years, for a Germany-wide average apartment of 65.14m² renovated to target 70 kWh/m²/y, with increasingly improbable economic conditions.

The third gives the percentage payback after 25 years, for an average large multi-apartment building in Germany renovated to 70 kWh/m2/y, again with a set of increasingly improbable scenarios that would be required in order to achieve 100% payback after 25 years.

Calculation of costs and benefits

I calculate costs and benefits in terms of their net present values (NPVs) as at the time of the renovations. If the NPV of the benefits is equal to or larger than that of the costs, the renovation is economically viable. The key equations used for this, given below, are integrated into a computer-based cost–benefit tool which is available for free download.¹⁰ For the full set of algorithms the reader may download the programming code, from http://justsolutions.eu/CostBenefits/.

Assuming some of the cost of renovation is financed through a loan, the monthly payment P on a loan of value L with interest rate R% per year for a term of n months is given by:

$$P=\frac{L \times {~}^{R}\!\left/ \!{~}_{1200}\right.\times {(1+{~}^{R}\!\left/ \!{~}_{1200}\right.)}^{n}}{{(1+{~}^{R}\!\left/ \!{~}_{1200}\right.)}^{n})-1}$$

(1)

The net present value M of loan repayments is then:

$$M=\frac{P\times (1-{F}^{n})}{1-F}$$

(2)

where F is a factor based on the discount rate:

$$F=1/(1+{~}^{D}\!\left/ \!{~}_{100}\right.)$$

(3)

and D is the monthly percentage discount rate factor based on an annual discount rate of b:

$$D= \sqrt[12]{1+b}-1$$

(4)

The amount of loan depends on the costs of energy-efficiency renovation plus PV installation cost (if any), less the property owner’s up-front cash investment in the project. This up-front investment incurs opportunity costs in the form of lost monthly income that could have been earned by an alternative investment. The opportunity costs have a net present value U, given by:

$$U=Q\times \frac{{D}_{U}}{1200}\times \frac{\left(1-{F}^{p}\right)}{\left(1-F\right)}$$

(5)

where Q is the amount of the property owner’s own up-front capital invested in the renovations and D_U is the expected annual rate of return for an alternative investment. Note that U is only the lost annual return, which counts as a cost in addition to the amount Q spent from savings.

The net present value K_npv of the total investment cost is therefore:

$${K}_{npv}=M+Q+U$$

(6)

We turn now to the benefits. The energy cost savings S_m in month m, as a result of the energy-efficiency upgrade are given by:

$${S}_{m}=\frac{\left({V}_{2}-{V}_{1}\right)\times A\times {C}_{m} }{12}$$

(7)

where V₁ and V₂ are the actual pre-and post-retrofit consumption (in German “Verbrauch”) in kWh/m²/y, A is the liveable floor area of the building in m², and C_m is the cost of energy per kWh in month m. The cost of energy C_m in a given month is given by:

$${C}_{m}={C}_{0}\bullet {\left(1+G\right)}^{m}$$

(8)

where C₀ is the initial energy cost and G is the monthly percentage inflation rate, given by:

$$G= \sqrt[12]{1+i}-1$$

(9)

where i is the annual inflation rate.

The Carbon tax savings S_Tm in month m are given by:

$$S_{Tm}=T_m\times{}^E/{{}}_{1000}\times\left(V_2-V_1\right)/12\times A$$

(10)

where T_m is the tax rate in month m in €/kgCO2 (the tax rate in €/tCO2 divided by 1000), and E is the CO2 emissions rate of the energy carrier in kgCO2/kWh, and A is the floor area. The tax rate varies according to the energy carrier and the building’s energy efficiency standard. A further set of algorithms is integrated into the tool to calculate this. The calculation of T_m has the same form as the calculation of Cm as in Eq. (8).

The net present value S_npv of 25 years’ worth of energy and tax savings, i.e., p months’ worth, is given by:

$${S}_{npv}= \frac{{(S}_{m}+{S}_{Tm})\times \left(1- {F}^{p}\right)}{1-F}$$

(11)

The main outputs the tool gives are the NPV of costs and benefits, the percentage NPV payback over the technical lifetime of the renovation measures, and the cost of abated CO2 (not discounted). The user can input the technical lifetime, the discount rate, inflation rates, mortgage interest rate and term, alternative investment rate of return, as well as the costs of the job and the pre-and post-renovation consumption. A screenshot of the tool’s user interface is given in Appendix 2 Fig. 5,  which also shows other outputs. These include, among other items: pre-and post-renovation CO2 emissions, monthly energy cost savings, the NPV of the cost of energy saved in €/kWh and, in case the property is rented, the maximum rent increase that would be cost-neutral for the tenants.

For the scenarios with rooftop photovoltaics, I normalised a year-long profile of photovoltaic electricity production at 15-min intervals (provided by the Fraunhofer Institute) to fit the size of the photovoltaic array. I assumed an annual yield of 950 kWh per kWp. I also normalised a year-long profile of typical electricity consumption of a household with a heat pump, to fit the sum of the sizes of the households in the apartment, again at 15-min intervals. I brought these together in a computer program to find, for each 15-min interval, the amount of photovoltaic electricity used by the apartments (including a portion for common areas) and the amount that would be fed into the grid. The cost of the photovoltaic systems was estimated from an actual installation of a 9.5 kWp system on a German house in 2022, together with component and installation prices in 2023. Heat pump performance is assumed to be in line with modelling done by Terry and Galvin (2023) for buildings renovated to a basic standard, namely an average annual coefficient of performance of 3.0 with a technical rebound of 20%.

Results

Initial analyses

The results of the initial analyses are given in Table 3, with rows and columns numbered for easy reference. Column 3 gives the target post-retrofit consumption, in kW/m²/y. The results indicate that none of the scenarios pay back. Columns 4 and 5 give the NPVs of costs and benefits, and in every case the costs exceed the benefits by a large margin. Column 6 gives the NPV of the losses (benefits minus costs), and column 7 gives these per m² of floor area, averaging 506 €/m² and ranging from 95 to 952 €/m². Column 8 gives the NPV of the cost of energy saved (€/kWh), which averages 0.422 and ranges from 0.144 €/kWh to 0.833 €/kWh, with median 0.27 €/kWh. To be economically viable, this value would need to be as low as or lower than the energy price, currently about 0.115 €/kWh for gas and 0.358 €/kWh for electricity.

Table 3 Main results of cost–benefit analyses

Column 12 gives what is probably the most telling result, showing the percentage of costs that are paid back, through energy savings plus CO2 tax savings, in 25 years, all in terms of NPV. This is also displayed in Fig. 1. The average payback is 22.5% over the 25 years, with maximum 53.7% and minimum 9.3%. Of the Germany-wide averages, the best result for the average-sized German apartment (Rows 21–24) is a 16.1% payback in 25 years; for the average small apartment building (Rows 34–39) it is 13.8% and for the average large apartment building (Rows 40–44) it is 16.3%. In no case, then, do any of these deep renovation scenarios come near to giving payback after 25 years.

Fig. 1

Percentage payback over 25 years, in terms of net-present values, of 44 case study energy efficiency renovations. "Anyway" costs not included. Target post-renovation consumption is given at end of each scenario label

The Aachen case scenarios are interesting because they utilize straightforward renovation to two different standards, then give scenarios where gas boilers are replaced with heat pumps, then with heat pumps plus shared rooftop photovoltaics. The basic standard of around 70 kWh/m²/y gives a return over 25 years of 29.08% (building 1) and 28.51% (building 2). The next highest standard of 50 kWh/m²/y gives a lower return, of 20.12% and 19.83%. Changing to heat pumps with the basic standard lowers the return even further, to 16.49% and 15.12% respectively. However, adding rooftop photovoltaics to this scenario increases the returns to 43.8% and 37.7% respectively. Note that, with heat pumps, the standard improves to 26 kWh/m²/y, since the amount of heating energy is reduced by almost one-third. As noted above, however, using a heat pump for a central heating system is only possible with a building that is already renovated to a standard of about 70 kWh/m²/y or better. The percentage paybacks are displayed graphically in Fig. 1.

Sensitivity analyses

The first sensitivity analysis, with results displayed in Fig. 2, shows the effect of having zero discount rate and zero opportunity costs, for a representative selection of the scenarios. This is equivalent to a simple back-of-the-envelope calculation. In all cases the percentage payback increases by about half its original value, but in no cases does it come near 100%. This shows that the failure to pay back is not just a fiction caused by the use of discount rates and opportunity costs.

Fig. 2

The effect on percentage payback time over 25 years, of zero discount rates and zero opportunity costs, for a selection of cases

The second sensitivity analysis, with results displayed in Fig. 3, gives the percentage payback in 25 years, for a Germany-wide average apartment of 65.14m² renovated to target 70 kWh/m²/y, with increasingly improbable favourable economic conditions. Case 1 gives the original conditions (discount rate 6%, etc., as shown in the labels on the chart). For case 2 the opportunity costs and opportunity costs are set at zero, and the percentage payback after 25 years increases from 9.3% to 17.1%. Case 3 extends Case 2 by increasing the energy and CO2 tax inflation rate from 2 to 4%, bringing an increase in payback to 22.5%. Case 4 extends Case 3 by reducing the own up-front capital to zero, and the payback reduces to 19.4%. Case 5 extends this by reducing the loan interest rate to 1%, and payback increases to 27.0%. Case 6 extends this by almost doubling the gas price to 0.21 €/kWh, thereby increasing payback to 48.2%. Case 7 extends this by increasing the initial CO2 tax from 30 to 600 €/tCO2, thereby increasing payback to 74.1%. In Case 8, a further increase in the energy and CO2 tax inflation rate to 6.1% drives payback to just over 100%.

Fig. 3

Sensitivity analysis: percentage payback in 25 years, for Germany-wide average apartment of 65.14m² renovated to target 70 kWh/m²/y, with increasingly improbable economic conditions

This illustrates that macroeconomic conditions, energy prices and CO2 tax would have to be absurd to achieve 100% payback – for example, near-zero interest rates together with an inflation rate of 6.8% that is ignored in discount rates, with energy and CO2 prices and inflation rates that would be impossibly high even for middle-income households.

The third sensitivity analysis considers the percentage return after 25 years, for a Germany-wide average large multi-apartment building renovated to 70 kWh/m²/y. The accompanying table, Table 4, gives the details of the scenarios. Scenario 1 is the basic case with 2% inflation, 6% discount rate, etc., as in Fig. 1. Scenario 2 shows that the percentage payback after 25 years increases from 11.5% to 20.9% if the discount rate and opportunity costs are set to zero. Scenario 3 builds on this, increasing the energy and CO2 tax inflation rate to 4%, which increases payback to 27.4%. Scenario 4 builds on this by increasing the initial energy price to 0.21 €/kWh, and payback increases to 48.9%. In scenario 5 the initial CO2 tax increases to 100 €/tCO2 and payback increases to 52.2%. Scenario 6 includes all these changes plus an increase in the energy and CO2 tax inflation rates to 8.5%, bringing a payback of 101% after 25 years. Once again we see that it is impossible to get anywhere near 100% payback without improbably, non-credibly high inflation rates, initial high energy and CO2 tax prices, and non- credibly low discount rates and opportunity costs. This is very important because this case represents an average of large apartment buildings in Germany (Fig. 4).

Table 4 Parameter values for 6 scenarios for third sensitivity analysis, based on for average large multi-apartment building in Germany renovated to 70 kWh/m²/y

Fig. 4

Percentage return after 25 years, for average large multi-apartment building in Germany renovated to 70 kWh/m²/y, with sensitivity analysis. Case scenario details in accompanying table

Discussion

What would be the implications of a finding that deep energy-efficiency renovation in Germany comes nowhere near to paying back through energy and tax savings within the technical lifetime of the renovation measures, for such a large and representative array of individual buildings and Germany-wide averages of types of buildings?

To begin with, this suggests that there is a missing or at least under-represented dimension within literature on the economics of energy efficiency. Interestingly a recent study of low-level, relatively inexpensive energy-efficiency improvements in buildings in the United States (Fowlie et al., 2018) also found that these do not pay back. Those authors note that: “Even when accounting for the broader societal benefits derived from emissions reductions, the costs still substantially outweigh the benefits.” They also comment: “The findings also suggest that the past several decades may have seen too much investigation into the why of the energy efficiency gap and not enough into whether there really was one.” A more nuanced response might be to suggest that literature on the energy efficiency gap would be greatly enriched by a slew of empirical studies on the economics of deep energy-efficiency renovation.

On the other hand, there are co-benefits of energy-efficiency renovation that are not reflected in a strictly economic analysis such as that presented here. For example, Baniassadi et al. (2022) found that “avoided health and climate costs could total around 40% of the direct utility savings”. If this is so, a case could be made for increasing the figure of percentage savings in the results of this study by 40% of their value. If we do this, the percentage payback then ranges from a minimum of 13.0% to a maximum of 75.18%, with an average of 31.5%%. This is still far short of economic viability, but it closes the gap somewhat. However, it also raises the question of whether or how society would pay this money, especially when the same case could be made for financial support for many other environmentally and socially valuable goals.

In relation to co-benefits, a case could be made for ignoring the prebound effect when doing cost–benefit analyses, as this would take account of the increase in thermal comfort due to deep renovation. Literature on prebound effects suggests that the healthy optimum is still lower than the official energy rating (Sun, 2014; Sun et al., 2014; Yan et al., 2017) and may be about halfway between this and actual consumption (Galvin, 2023a, 2023b). In this study’s analysis, cases 25–33 deliberately ignore prebound effects, i.e. they use the official energy rating for the pre-renovation consumption. However, the payback after 25 years for these scenarios ranges from 10.9% to 20.7%, with average 14.9%. This is not very promising.

For a Germany-wide average large apartment building, consider the best performing renovation scenario (scenario 43). The percentage payback after 25 years is 16.3% using actual pre-retrofit consumption of 114 kWh/m²/y, and increases to 23.3% with an average theoretical pre-renovation consumption of 147 kWh/m²/y, i.e., when the prebound effect is ignored. This makes a difference, but it does not bring the scenario close to payback, and in all these cases the owner-occupiers would have to find extra money to pay for the energy that they are not actually saving.

Regarding the sales market value of a building, there is abundant evidence that higher energy efficiency increases this (Taruttis & Weber, 2022. As Galvin (2023a) shows, however, there can be three different consequences of this in Germany. First, a building owner who sells immediately after renovating is likely to get the money back which they paid for energy-efficiency renovations, though this varies regionally. Second, a building owner who lives in the property still suffers the loss due to energy and tax savings not giving a full payback. Third, a purchaser of the property suffers the same loss. The problem is that the sums do not add up to a full payback, so at least one of the actors loses out.

Regarding society, there is the benefit of reduced fine-particle emissions due to reduced gas consumption or the replacement of gas with electrically driven heat pumps (Baniassadi et al., 2022), and the question again arises as to who should pay for this.

Society also benefits from reduced CO2 emissions, but this is not straightforward. Column 9 of Table 3 gives the costs of CO2 saved, in €/tCO2, for each of the scenarios. This ranges from 677 €/tCO2 to 3,578 €/tCO2, with median 1,964 €/tCO2. Most values range between 1,000 €/tCO2 and 3,000 €/tCO2. This seems an extraordinarily high price to pay, when we can reduce CO2 emissions much more cheaply by other means. For example, Misconel et al. (2022) estimate costs of CO2 abatement at around 50–100 €/tCO2 for onshore wind, < 40 €/tCO2 for ground-mounted photovoltaics, and 160–1,300 €/tCO2 for rooftop photovoltaics, depending on the situation. Policymakers need to give more thought to how best to direct the costs of CO2 abatement. On the one hand, if all homes switched to heat pumps and the grid were decarbonised, this might reduce the cost of CO2 abatement deeply. On the other hand, it is not yet clear whether a fully decarbonised grid is possible, so there may well be an argument for driving energy efficiency levels higher despite the excessive costs per tonne of CO2 abated and per kWh of energy saved. In any case, however, if we want to call these renovations “economically viable” we need to be very careful to define precisely what we mean.

Conclusions

This paper investigated whether deep energy efficiency renovation of Germany’s older residential buildings pays back, in monetary terms, over a 25-year technical lifetime of the renovation measures. It did so by way of cost–benefit analyses of 44 case study scenarios, comparing net-present values (NPVs) of costs with NPVs of benefits. These included 29 scenarios dealing with specific buildings, and 15 based on Germany-wide averages of types of buildings: the average sized German apartment, the average “small” apartment building (< = 5 apartments) and the average “large” apartment building (> 5 apartments). “Deep” renovation aims to achieve a space heating energy efficiency level sufficient to enable Germany to reach its climate goals, namely about 70 kWh/m²/y or better. It involves expensive elements such as external wall insulation and window replacement. Further, simply installing an air-to-water heat pump without deep building envelope renovation will not achieve the goal of 70 kWh/m²/y because the water temperature in the radiators will be insufficient to provide adequate space heating without a drastic fall in heat pump efficiency.

The investigation was set alongside a narrative in economics literature that energy efficiency upgrades are economically viable, and the specific claim by the German government and a number of interest groups that deep energy-efficiency renovation of Germany’s old buildings to the specified minimum standard is always economically viable. The study found that, on average, only 22.5% of the costs paid back, through energy and CO2 tax savings, after 25 years of operation, with a range of 9.7% to 53.7%.

The study used typical values for discount rates, opportunity costs, loan interest rates, energy price and CO2 tax inflation, the current CO2 tax rate for space heating, and average energy prices based on over 100 energy providers’ charges. However, it also performed sensitivity analyses in which the values of these parameters were changed to enable higher percentage payback. In all cases, 100% payback could only be achieved with zero discount rates, zero opportunity costs and improbably high energy prices, CO2 taxes and inflation rates.

The study also tested a key societal benefit of deep renovation, namely CO2 emissions reduction. It found that the cost of CO2 abatement through deep renovation was an order of magnitude higher than straightforward electricity grid decarbonisation through wind power or photovoltaics, ranging from 677 €/tCO2 to 3,578 €/tCO2, with an average of 1,963 €/tCO2.

Nevertheless, this does not imply that we should abandon deep energy efficiency renovation. Instead, we can benefit from seeing that property owners are not always avoiding it due to economic irrationality. Their hesitations due to its percentage monetary payback need to be taken seriously. To increase the rate of CO2 abatement we need to start with the best information possible, not with a doctrinaire assertion that “energy-efficiency upgrades always pay back”. If it is not possible to upgrade the building stock economically viably, we need to think about what this means. For example, it could be that we gradually transition to heat pumps and shift the problem of achieving net-zero-CO2 to the electricity grid. On the other hand, transitioning the grid to net-zero might not be materially possible – but again, it is important to face the actual challenges, rather than be guided by a poorly grounded, somewhat doctrinaire assumption.

Appendices

Appendix 1. The “anyway” costs

In estimating the cost of energy-efficiency renovation, German policymakers do not include the costs of work and components that would be incurred if an old building were being brought up to a good standard without energy-efficiency improvements. These are called “anyway” costs (ohnehin- or umsonst- or sowieso-Kosten). A typical definition of these costs, as given by the Institut Wohnen und Umwelt (http://www.iwu.de) is outlined in Appendix Table 5.

Table 5 “Anyway” costs as defined by the Institut Wohnen und Umwelt, as given in Hinz and Enseling (2021), P. 26

Appendix 2. The computer tool used for the cost–benefit analyses.

Fig. 5

Screenshot of downloadable tool used for cost–benefit analyses, designed by the author. Free to download from www.justsolutions.eu/TenantCBA/OwnerOccCBA_PV_option_Vs11_Beta_Package17Feb2024.zip