Energy efficiency improvements: implications for carbon emissions and economic output in Germany

In order to combat climate change, Germany has committed to the goal of significantly reducing its greenhouse gas emissions by 2030. It is generally agreed that energy consumption will have to be lowered to achieve this, besides switching to low-carbon energy sources. Carbon pricing is a key instrument used in climate policy to better account for the climate costs entailed by consumption of fossil fuels and to set incentives to limit this consumption. Increases in energy efficiency could help to reduce energy consumption without any shortfalls in aggregate output. Energy efficiency improvements mean that, all else being equal, the same quantity can be produced using less energy or that more can be produced using the same amount of energy as before.

Over the past three decades, aggregate energy efficiency in Germany, as measured by energy-saving technological progress, has risen by an estimated 2.8 % per year. However, the increases in energy efficiency have varied greatly across different economic sectors.

Using model simulations, it is possible to gauge the extent to which energy efficiency improvements can reduce emissions and boost economic activity. The approach can also be used to simulate the role of the carbon price in production costs and the associated dampening of activity and lowering of emissions. Ambitious carbon pricing is presumed to also create incentive effects for energy efficiency improvements, thereby leading to greater reductions in emissions. This, in turn, could feed back into a dampening effect on the carbon price. The existing analytical framework cannot yet depict these kinds of important interactions between the carbon price and energy efficiency improvements. However, they would not fundamentally change the core messages of the analysis.

The analysis shows that energy efficiency improvements lower aggregate energy intensity and boost aggregate output. They also counteract activity-dampening contributions, which are likely to stem, at least for a time, from the rising price of carbon emissions. Greater energy-saving technological progress would also facilitate stricter climate policy, such as a higher carbon price. Energy efficiency improvements are thus an important factor in achieving climate objectives.

Without changes to the planned carbon price pathway and given a continuation of the energy efficiency improvements of the past thirty years, simulation results show that the reduction in emissions will not meet the targets set by politicians. These analyses indicate that there would have to be a considerable increase in energy-saving technological progress, also in response to the climate action taken, for the planned reduction in emissions to be achieved.

It is also evident that the emission targets can be met, for example, by raising the carbon price more strongly. The extent to which this is accompanied by output losses or gains depends, amongst other things, on the strength of energy-saving technological progress. A higher carbon price is likely to boost incentives for such progress.

1 Energy efficiency in the context of climate policy

Germany is taking several routes to achieve the planned reductions in greenhouse gas emissions. According to the Federal Climate Change Act (Bundesklimaschutzgesetz), Germany still needs to reduce its greenhouse gas emissions by around 35 % by 2030 compared with 2023 levels. 1 This emission reduction target goes beyond the European Commission’s objectives for the EU. 2 Key instruments used to curb greenhouse gas emissions are taxation of emissions (carbon price) and investment in renewable energies and low-carbon technologies. Improvements in energy efficiency aid in achieving climate objectives (International Energy Agency (2023), Intergovernmental Panel on Climate Change (2018)). Increases in energy efficiency make it possible to achieve a given level of production using less energy, or a higher level of production at a given amount of energy use. 3 The decisions made at the 28th United Nations Climate Change Conference (COP28) in Dubai in December 2023 also highlighted this aspect.

Carbon pricing is a useful tool for effectively reducing emissions. Viewed in isolation, carbon pricing increases the costs of using fossil energy under otherwise identical conditions. This can lead to output losses in the short term. 4 In the long term, however, the transition to global net zero carbon emissions will likely lead to higher aggregate output than a scenario in which no climate action is taken, as potential losses caused by climate change would be avoided. 5

Energy efficiency improvements are directly linked to lower greenhouse gas emissions via energy consumption. Using energy more efficiently reduces energy consumption and thus emissions, all else being equal. This makes energy efficiency a key factor in cutting greenhouse gas emissions. Improvements in energy efficiency draw attention to an additional component: energy-saving technological progress. As this boosts GDP growth, it plays an important role in the transition to a low-carbon economy.

There are interactions between carbon pricing and energy-saving technological progress that cannot be directly depicted in the chosen model framework as yet. As a higher carbon price tends to make energy use more expensive, firms faced with rising carbon prices have a greater interest in developing and deploying energy-saving technologies more rapidly. 6 This increases the effect of carbon pricing on emission reductions. 7 At the same time, a stronger reduction in emissions could have feedback effects on the carbon price. However, empirical evidence on the degree to which carbon pricing affects energy efficiency improvements is scarce. 8 The following analysis uses a model framework in which energy efficiency gains are exogenous. In the model, they are not directly influenced by macroeconomic developments or economic policy measures such as a carbon price. Instead, assumption-based scenarios are defined, setting out different pathways for energy efficiency and emissions pricing. This depicts the two driving forces of emission reduction more clearly in isolation, but disregards presumably important interactions for now.

The text below analyses how energy efficiency improvements in conjunction with the carbon price pathway can help reduce emissions by 2030 and what macroeconomic consequences this will have. Simulations using the environmental multi-sector DSGE model developed at the Bundesbank – EMuSe – can provide insights here. 9 To do this, the model is fed with both the currently foreseeable increase in the cost of fossil energy given the planned carbon price pathway as well as energy efficiency improvements in line with past developments. 10 Simulated changes in both pathways make it possible to estimate how much each of them contributes to reducing emissions by 2030 and what impact they have on aggregate output up until then.

In particular, there is the question of whether Germany’s emission reduction targets can be achieved with the planned carbon price pathway given an unchanged pathway for energy efficiency improvements. The carbon price pathways of both the national and the EU emissions trading systems contribute to reducing emissions. What is unknown is how close emissions will get to their target level given the status quo of the carbon price pathways. Alternative scenarios can be used to estimate the adjustments to the emission price or the additional improvements in energy efficiency that might be necessary to achieve the emission reduction targets. This also gives an indication of how strong the incentive effects of climate action, such as the carbon price, would need to be to generate the energy efficiency improvements necessary for the required reduction in emissions.

2 The importance of energy efficiency and how it relates to energy intensity

The link between emission reductions and lower energy consumption is fairly clear-cut. As greenhouse gas emissions are generated by the burning of fossil fuels, it follows that energy consumption from fossil fuels has to drop in order to reduce emissions. Improvements in energy efficiency lower the energy consumption required for a given level of production. This would mean that lower emissions could be achieved without any output losses. 11 12

Viewed from the angle of production theory, the term energy efficiency improvements refers to energy-saving technological progress. Energy efficiency feeds into a conventional production function as a technological variable, similar to total factor productivity. However, this form of technology is linked to the use of energy. For firms, energy-saving technological progress means that they can use energy more efficiently in their production. Examples of this are new machines that deliver the same performance but require less power, or recycling the heat generated when burning fossil fuels for use in production processes. In these cases, primary energy consumption is thus reduced for a given level of production. The main focus of energy efficiency in the sense of technological progress is therefore on how effectively energy is used in production.

From a macroeconomic perspective, energy efficiency often refers to energy intensity, which is the ratio of aggregate energy consumption to GDP. In particular, the COP28 decisions on energy efficiency are based on energy intensity. In this case, an increase in energy efficiency means a lower energy intensity. Consequently, it is not solely about how effectively energy can be used, but more generally about reducing energy consumption relative to output. The source of this shift is not relevant for determining energy intensity. Therefore, this is not a technological variable, but an observable measure of energy efficiency.

Changes in aggregate energy intensity do not necessarily reflect changes in energy efficiency in the sense of technological progress. For example, sectoral shifts may lead to a reduction in aggregate energy intensity without any efficiency gains in the use of energy being generated within production processes. This occurs, for instance, when the weight of the services sector increases relative to more energy-intensive economic sectors. Another factor that can lead to a reduction in energy intensity through lower energy use is the substitution of energy for other goods in the production process. 13 Even so, energy efficiency in the sense of energy-saving technology is closely linked to energy intensity. Energy-saving technological progress means less energy being used for the same level of production. Energy intensity therefore decreases. Thus, energy-saving technological progress can lead to lower energy intensity, but a reduction in energy intensity does not necessarily require an increase in energy-saving technology.

3 Energy efficiency improvements in Germany

To estimate the magnitude of potential future energy efficiency improvements, it helps to look at their evolution over the past decades. This gives indications of the pace of energy efficiency improvements in different sectors of the economy that could be achieved in the future based on past developments.

A production theory approach can be used to determine the energy efficiency improvements made over the past three decades in the sense of energy-saving technological progress. By formulating a production function incorporating technological progress specific to energy, it is possible to fit the available data into a structure in order to calculate energy-saving technological progress (see the supplementary information on measuring energy-saving technological progress). 14 Energy efficiency improvements can thus also be determined for different sectors of an economy. On that basis, both the macroeconomic trend and sectoral specificities can be analysed. Here, the energy efficiency gains for ten sectors (based on the European classification of economic activities) for the years 1991 to 2019 are calculated. 15

Supplementary information

Measuring energy-saving technological progress

Improvements in energy-saving technology in the production process are a key driver of aggregate energy savings. The level of energy-saving technology denotes how efficiently energy is used in the production process. Energy-saving technology is not an observable variable, which is why it has to be estimated. Quantifying energy-saving technology and how it changes over time is key to determining the extent to which it contributes to the evolution of output and emissions.

Production theory can be used as an approach for computing energy-saving technology. 1 According to a theoretical production function, factors of production are combined to produce output. The impact of using the individual factors of production as inputs depends in part on how efficient they are. To estimate improvements in energy efficiency at the sectoral level, the sectoral production structure of EMuSe is used in the calculations presented below. In the EMuSe model framework, production ($ y_{s,t} $) in sector $ s $ at time $ t $ is a function of the input factors labour ($ N_{s,t} $) and capital ($ K_{s,t} $) as well as the intermediate inputs ($ H_{s,t} $):

$$ y_{s,t} = \left(K_{s,t}^{1 – \alpha_{N,s}} N_{s,t}^{\alpha_{N,s}} \right)^{1 – \alpha_{H,s}} H_{s,t}^{\alpha_{H,s}} $$

The intermediate inputs consist of energy ($ E_{s,t} $) and non-energy intermediate inputs ($ NE_{s,t} $):

$$ H_{s,t} = \left[\alpha_{NE,s}^{1 – \sigma_H} NE_{s,t}^{\sigma_H} + \left(1 – \alpha_{NE,s}\right)^{1 – \sigma_H} \left(\mathbf{\epsilon_{s,t}} E_{s,t}\right)^{\sigma_H}\right]^{\frac{1}{\sigma_H}} $$

The parameters $ \alpha_{N,s} $, $ \alpha_{H,s} $ and $ \alpha_{NE,s} $ in both equations provide information on how sensitive production is to changes in the individual factors of production. 2 The parameter $ \sigma_{H} $ indicates the degree of substitutability between energy and non-energy intermediate inputs. 3 The variable $ \epsilon_{s,t} $ describes the level of technology that influences the efficiency of energy in the production process. Here, an increase in $ \epsilon_{s,t} $ can be interpreted as energy-saving technological progress.

Building on the production theory approach, observable variables can be used to draw conclusions about unobservable energy-saving technology. According to this approach, a correlation exists between energy-saving technology and relative developments in intermediate inputs and energy consumption, as well as the share of energy costs in output. It is assumed that perfect competition exists in product and factor markets. 4 If price-adjusted energy consumption rises more slowly than price-adjusted intermediate inputs, all else being equal this indicates an increase in energy-saving technology, meaning more efficient use of energy in the production process. In addition, a higher share of energy costs in gross output, according to the model, is associated with a decline in energy-saving technology. 5 A decline of that kind can come about, for example, as a result of organisational changes within enterprises that lower the efficiency of energy input in production.

Results for developments in energy-saving technology between 1991 and 2019 can be calculated based on sectoral data from the Federal Statistical Office. Data on gross output and intermediate inputs (including the associated price indices) are taken from the national accounts. Data on energy costs are sourced from the input-output tables. The Environmental Economic Accounts provide data on sectoral energy consumption. The sectoral energy price (cost per unit of energy consumed) is calculated by expressing sector-specific energy costs in relation to energy consumption. These data can be used to calculate energy-saving technology for the years 1991 to 2019. 6

There are other methods besides the approach based on model or production theory for measuring changes in energy efficiency. In the empirical literature, energy efficiency shocks are also identified using structural vector autoregressions (SVAR). 7 Jo and Karnizova (2021), for example, interpret energy efficiency gains as economic shocks that induce the opposite effects on GDP and emissions. 8 Unlike the approach based on production theory, however, this analytical framework does not allow historical trends in energy efficiency to be investigated. Moreover, the energy efficiency gains calculated using the production theory approach are consistent with the production structure in the EMuSe model that is used in the main article to analyse energy efficiency improvements.

Footnotes

See Hassler et al. (2021). The Joint Economic Forecast (2022, 2023) also calculates energy-saving technology for Germany using a single-sector model that has a different production structure than EMuSe. Energy-saving technology can be calculated as a residual, similar to the Solow residual. This can be done using the factor demand functions for energy and intermediate inputs in general. These equations derive from the optimisation problem of enterprises.
The parameter $ \alpha_{H,s} $ denotes the production elasticity of the intermediate inputs and the long-term share of production factor $ H $ in output in sector $ s $. The production elasticity of labour, by contrast, is indicated by $ (1 – \alpha_{H,s})\alpha_{N,s} $ and the long-term share of labour as a factor of output in sector $ s $. The parameter $ \alpha_{NE,s} $, meanwhile, stands for the long-term share of non-energy intermediate inputs in the total intermediate inputs used in sector $ s $.
For negative values of $ \sigma_{H} $, energy and non-energy intermediate inputs have a low level of substitutability in production; positive values mean that they are readily substitutable.
Perfect competition in factor and product markets is a standard assumption made as part of a medium-term analysis of the effects of technological progress on the real economy. See, for example, Hassler et al. (2021).
In the calculations, it is assumed that $ \sigma_{H} = − 9 $, based on insights from the empirical literature (Atalay (2017), Barrot and Sauvagnat (2016) and Boehm et al. (2019)).
The input-output tables are available up to 2020. As 2020 is strongly influenced by pandemic effects, energy-saving technology is only calculated until 2019 in this case.
See Bruns et al. (2021) and Jo and Karnizova (2021).
This definition is not clear-cut, as it is possible to lower emissions without burdening GDP even in the absence of energy efficiency gains. This is the case, for example, when the energy mix shifts towards lower-emission or non-fossil-based energy. Bruns et al. (2021) feed energy consumption into the estimation as a variable to obtain a more reliable estimate of energy efficiency shocks. Their analysis does not, however, include emissions in the estimation as an additional variable, which means that no conclusions can be drawn regarding the impact of energy efficiency gains on emissions.

Overall, energy efficiency has risen continuously over the past three decades, but there have sometimes been considerable differences across sectors. On balance, the level of energy-saving technology in Germany has more than doubled over this period. On aggregate, it increased by just under 2.8 % per year. 16 Across the various sectors of the economy, the annual increase varied between ½ % and 5 % from 1991 to 2019. The largest improvements in energy efficiency were achieved in the water supply sector, but the other services sector and the electricity and gas supply sector also significantly increased their energy efficiency. By contrast, energy efficiency gains were small in the transportation and storage sector and in the fossil energy sector. Energy-saving technological progress in the fossil energy sector has actually declined since 2012. In addition, the ranking also changed over time. While the transportation and storage sector recorded strong energy efficiency gains in the late 1990s compared with the other services and manufacturing sectors, the latter two sectors have overtaken it since 2014.

Advancements in energy-saving technology were not reflected to the same extent in reduced energy intensity. 17 Between 1991 and 2019, energy intensity in Germany declined steadily, but it only decreased by just under 2 % per year. One reason for this is probably rebound effects. 18 Overall, there was a reduction of 40 %.

4 Importance of energy efficiency as shown by the EMuSe model

The importance of energy efficiency can generally be analysed using various types of models. On the one hand, empirical approaches such as structural vector autoregressions (SVAR) can be used to determine the macroeconomic impact of energy efficiency improvements. 19 One advantage of these approaches is that they do not require assumptions about a specific production structure. However, certain assumptions about the relationship between the underlying drivers are necessary in this case, too, so that statements about the macroeconomic effects can be made. On the other hand, theoretical model approaches such as dynamic stochastic general equilibrium (DSGE) models are an option. The main advantage of this type of model over purely econometric estimation methods is that its structure allows it to shed light on different transmission channels for the results. In terms of our research question, these model types allow us to look at other specific factors such as sectoral heterogeneity, interlinkages in production or emissions, besides the possible analysis of trend developments. 20 They are also able to run mechanisms that were weak or absent in the past. This is particularly relevant for the impact of the carbon price, which has been significantly higher especially since 2021.

The EMuSe model developed at the Bundesbank is an environmental multi-sector DSGE model. EMuSe can be used to analyse climate policy adjustment processes in particular, and the model can be adapted flexibly to suit the respective research purpose. EMuSe was developed specifically to enable comparatively detailed analyses of the interactions between the economy and climate policy over a timeframe relevant for monetary policy. This is typically much shorter than that in conventional climate impact models with macroeconomic variables, such as integrated assessment models (IAMs). 21

A special feature of the EMuSe model is that it contains linkages between the production in different economic sectors. Firms use not only capital and labour in the EMuSe model, but also intermediate inputs to produce output. These can come from any sector, although the extent to which various inputs are substitutable is limited. The composition of the intermediate input bundles varies depending on the sector. If, for example, the carbon price is not levied in all sectors at the same time as is currently planned, or if there are greater technological advances in a particular sector, this will have an impact on more than just the sector in question. Given that goods flow into other sectors as intermediate inputs, this is transferred to other sectors through relative price changes. In addition, the production structure contains sector-specific energy and non-energy intermediate inputs. Energy-saving technological progress is thus also sector-specific. Furthermore, the model depicts the special role played by energy in production, especially on account of its low degree of substitutability with non-energy intermediate inputs. 22 The model also includes greenhouse gas emissions generated by using fossil energy. 23

Model simulations using EMuSe provide scenarios for how emissions and production might evolve up to 2030 against a backdrop of different calibrations of the carbon price pathway and improvements in energy efficiency. 24 These simulations illustrate developments in emission and output levels that the model shows could arise from the interplay between energy efficiency gains, extrapolated from historical trends, and carbon pricing. Furthermore, it is possible to gauge how the carbon price pathway and energy efficiency improvements contribute to achieving the defined climate objectives.

Note that within the model framework energy-saving technological progress is exogenous. In the model, it cannot be attributed to relative price changes or economic policy instruments, such as spending on research and development. This also applies to carbon pricing, which does not influence the development of energy-saving technology in the model framework chosen here. Fundamentally speaking, incentives to improve energy efficiency and the allocation of resources towards achieving that goal can plausibly be expected to increase as the carbon price rises. That means the carbon price’s impact in terms of reducing emissions is likely to be underestimated and probably represents the lower bound of the effects that can be anticipated. On the other hand, it should be noted that the efficiency improvements fed into the current analysis are cost-neutral. 25 This relationship is not modelled in detail here, partly because it is difficult to estimate how and with what lags expenditure on research and development will be converted into energy-saving technologies. 26 One advantage of the simplified modelling method used here is that it is easier to determine how the aggregate impact of an increase in the price of carbon emissions can be changed by energy efficiency improvements. In addition, the respective pull forces can be better distinguished. 27 Economic policy measures other than the carbon price, such as an increased focus on expanding renewable energy sources, are disregarded in the model.

The starting point for these considerations is a hypothetical situation without any energy efficiency improvements. This scenario provides the basis for distinguishing the contribution the carbon price pathway makes to developments in production and emission levels from that made by energy efficiency gains. The carbon price pathway in this model variant corresponds to the developments in the carbon price planned from today’s perspective. The carbon price pathway applies in the model to the transportation and storage sector, the fossil energy sector and the energy-intensive manufacturing sectors. This reflects the actual coverage of sectors that are required to pay a carbon price either under the national or the EU emissions trading system (EU ETS). 28 In the model, revenue from carbon pricing is returned directly to households on a lump-sum basis. It should be noted that the simulations only examine how the carbon pathway contributes to the evolution of emissions and aggregate output. Other factors, such as developments in total factor productivity or energy efficiency improvements due to the higher carbon price, are disregarded.

According to the simulations, the reduction in emissions as a result of today’s planned carbon price pathway alone is not sufficient to drive down emissions and energy intensity to the degree aspired to. The model suggests a 15.7 % reduction in emissions by 2030 relative to 2023. However, the Federal Climate Change Act stipulates that a further reduction in emissions of 19.2 % relative to 2023 would need to be made. In addition, production becomes more expensive in those sectors where the emission price rises. Sectoral interlinkages also push up production costs for other sectors. Thus, compared with a reference scenario without carbon pricing, the reduction in emissions is associated with a decline in output. 29 However, the activity-dampening effect of carbon pricing is small. Overall, aggregate output falls by just 1 % by 2030 (all other things being equal). 30 Since the emission price curbs energy consumption and production to a similar extent, energy intensity remains virtually unchanged. Therefore, contrary to what is envisaged in the COP28 decisions, energy intensity in this scenario does not decline more sharply than has thus far been the case. These findings by no means provide an argument against carbon pricing – in the model it is certainly an effective way to reduce emissions. 31 According to model simulations, however, today’s planned carbon price pathway is not sufficient in itself to achieve the climate objectives. Moreover, it is evident that aggregate energy intensity will hardly decrease if there are no additional improvements in energy efficiency.

In a second scenario, energy efficiency gains are added to the model in addition to the carbon price pathway planned from today’s perspective. This scenario (status quo scenario) is based on expected developments in the carbon price and a continuation of the energy efficiency improvements achieved to date. The results are therefore to be interpreted as the contributions of these two variables to aggregate developments in emission and output levels and do not constitute a GDP forecast. The assumed rate of increase in energy efficiency in the economic sectors corresponds to the average sectoral rate of increase in energy-saving technology between 1991 and 2019. 32 As energy-saving technological progress is assumed to be exogenous, these energy efficiency gains are cost-neutral. In these model simulations, most sectors increase their production and reduce the use of energy. Output grows but emissions fall due to declining energy usage. This is ensured by the energy efficiency improvements. In the short term, the dampening impact of the carbon price pathway on production predominates, but the contribution that improved energy efficiency makes to growth continues to increase over time and output is around 1½ % higher in 2030 than in the absence of carbon pricing and energy efficiency gains. As such, the transition to a low-carbon economy triggered by carbon pricing is able to boost activity, provided it is accompanied by improvements in energy efficiency. Since energy efficiency increases as well as the carbon price pathway, emissions fall more sharply than in the absence of the contribution made by efficiency improvements. 33

The reduction in emissions in the status quo scenario also does not come close to the targets set out in the Federal Climate Change Act. Emissions in this scenario fall by close to 26.3 %. Thus, even in this scenario, a gap of almost 9 percentage points remains compared with the reduction in emissions meant to be achieved by 2030. Energy intensity decreases considerably in the status quo scenario, but falls short of the figure aimed for. It declines by an average of 2 % per year in the period up to 2030. The reduction is thus below the global target of 4 % set by COP28.

According to these model calculations, a stronger increase in the carbon price than is currently planned would therefore be necessary to achieve the emission targets outlined in the Federal Climate Change Act. Carbon pricing is very effective in reducing emissions in the model. However, even if energy efficiency improvements continue to follow the existing trend and are factored in to the calculation, the carbon price pathway would have to increase twofold over the entire simulation period. 34 According to the model simulations, such an increase in the carbon price pathway, when viewed in isolation, would lead to output losses for a certain period of time. However, rising energy efficiency offsets this. It makes energy use cheaper and production recovers considerably despite the rather steep carbon price pathway, meaning it stands at a slightly higher level in 2030 than in the absence of carbon pricing and energy efficiency gains. Aggregate energy intensity hardly decreases compared with the status quo scenario, however. This is because, here, too, it is assumed that the improvement in efficiency stems from the same energy-saving technological advances as before. The steeper carbon price pathway has an equally dampening effect on energy consumption and production and therefore does not lead to a further reduction in energy intensity.

If the carbon price pathway stays as currently planned, advances in energy-saving technology would need to be significantly stepped up in order to achieve target emissions. With carbon pricing on its current path, the rate of progress would have to increase by almost two-and-a-half times in each sector. Progress of this magnitude has been made in individual years and sectors over the past three decades. However, it is questionable whether the carbon price pathway envisaged thus far will generate sufficiently strong incentives to ensure that such improvements in efficiency are achieved across the board and on a sustained basis. There remains much analysis to be done in this regard.

5 Conclusion

According to the model results, improvements in energy efficiency are a key factor in achieving climate objectives without any large output losses. Increased energy efficiency lowers aggregate energy intensity and boosts aggregate output. As a result, energy efficiency improvements are an important factor – albeit one that cannot be directly controlled – in mitigating potential adverse macroeconomic effects caused by other climate policy measures. There is a two-way relationship between greater technical advances in the area of energy saving and a stricter climate policy – a higher carbon price than is currently envisaged, for instance. This is disregarded here, meaning that historical developments might tend to represent a lower bound for the potential of efficiency gains.

As demonstrated by the model, a combination of increased energy efficiency and strict climate action makes it easier to achieve the emission targets. However, simulation results suggest that energy-saving technology would need to improve significantly for the planned reduction in emissions to be achieved through this progress alone. It seems questionable whether the carbon price pathway currently envisaged already creates the necessary incentives for this. According to the model, if these incentives are not sufficient, further measures such as a steeper carbon price pathway would be needed to achieve the climate objectives. A higher carbon price, accompanied by greater energy-saving technological advances, could also entail growth in output over the course of time. Energy efficiency improvements could also be stimulated by stepping up efforts to promote research and development in this field. Alongside efficiency gains, other factors, such as the expansion of renewable energy sources, networks and storage facilities, are key for achieving the emission targets. According to the Federal Government’s current greenhouse gas projections, the climate objectives for 2030 are achievable, although a much faster pace for the expansion of renewable energy sources plays an important role. 35 The success of such measures also depends heavily on whether consistent measures are implemented and pursued alongside the objectives. As part of the energy transition and climate policy, policymakers are, in particular, tasked with setting reliable framework conditions.

6 List of references

ADAC (2024), Spritpreis-Entwicklung: Benzin- und Dieselpreise seit 1950, accessed on 2 January 2024.

Acemoglu, D., P. Aghion, L. Bursztyn and D. Hemous (2012), The Environment and Directed Technical Change, American Economic Review, Vol. 102, No 1, pp. 131‑166.

Aghion, P., A. Dechezleprêtre, D. Hémous, R. Martin and J. Van Reenen (2016), Carbon Taxes, Path Dependency, and Directed Technical Change: Evidence from the Auto Industry, Journal of Political Economy, Vol. 124 No 1, pp.1‑51.

Andrés J., J. E. Bosca, R. Doménech and J. Ferri (2024), Transitioning to Net-Zero: Macroeconomic Implications and Welfare Assessment, accessed January 2024.

Atalay, E. (2017), How important are sectoral shocks?, American Economic Journal: Macroeconomics, Vol. 9, No 4, pp. 254‑280.

Barrot, J.-N., and J. Sauvagnat, (2016), Input Specificity and the Propagation of Idiosyncratic Shocks in Production Networks, The Quarterly Journal of Economics, Vol. 131, No 3, pp. 1543‑1592.

Boehm, C. E., A. Flaaen and N. Pandalai-Nayar (2019), Input Linkages and the Transmission of Shocks: Firm-level Evidence from the 2011 Tōhoku Earthquake, Review of Economics and Statistics, Vol. 101, No 1, pp. 60‑75.

Bönke T., G. Dany-Knedlik and W. Roeger (2023), Erfüllung der Klimaziele kann nur bei richtiger Kombination der Maßnahmen Wachstumsimpulse geben, DIW Weekly Report No 34+35/2023, pp. 453‑468.

Brand, C., G. Coenen, J. Hutchinson and A. Saint Guilhem (2023), How will higher carbon prices affect growth and inflation?, The ECB Blog, 25 May 2023.

Bruns, S. B., A. Moneta, and D. I. Stern (2021), Estimating the economy-wide rebound effect using empirically identified structural vector autoregressions, Energy Economics, Vol. 97, No 105158.

D’Arcangelo, F. M., G. Pavan and S. Calligaris (2022), The Impact of the European Carbon Market on Firm Productivity: Evidence from Italian Manufacturing Firms. FEEM Working Paper No 24.

Deutsche Bundesbank (2022), Climate change and climate policy: analytical requirements and options from a central bank perspective, Monthly Report, January 2022, pp. 33‑61.

Federal Environment Agency (2024), Treibhausgas-Projektionen 2024 – Ergebnisse kompakt, accessed March 2024.

Federal Environment Agency (2023), Der Europäische Emissionshandel, price developments of EU emission allowances (EUA) since 2008, accessed on 23 November 2023.

Federal Government (2023), Ein Plan fürs Klima, information on the revised version of the Federal Climate Change Act (Klimaschutzgesetz), accessed on 21 June 2023.

Federal Ministry of Finance (2022), Fragen und Antworten zum Klimaschutz, accessed on 30 August 2022.

Hassler, J., P. Krusell and C. Olovsson (2021), Directed technical change as a response to natural-resource scarcity, Journal of Political Economy, Vol. 129, No 11, pp. 3039‑3072.

Hinterlang, N., A. Martin, O. Röhe, N. Stähler and J. Strobel (2023), The Environmental Multi-Sector DSGE model EMuSe: A technical documentation, Bundesbank Technical Paper No 03/2023.

Hinterlang, N., A. Martin, O. Röhe, N. Stähler and J. Strobel (2022), Using energy and emissions taxation to finance labor tax reductions in a multi-sector economy, Energy Economics, Vol. 115, No 106381.

Hulten, C. R. (2001), Total Factor Productivity: A Short Biography, in: New Developments in Productivity Analysis, pp. 1‑54.

Intergovernmental Panel on Climate Change, IPCC (2018), Global Warming of 1.5°C, Special Report

International Energy Agency (2023), Energy Efficiency 2023, IEA.

Jo, S. and L. Karnizova (2021), Energy Efficiency and Fluctuations in CO2 Emissions, Bank of Canada Staff Working Paper No 2021‑47.

Joint Economic Forecast (2023), Gemeinschaftsdiagnose Frühjahr 2023: Inflation im Kern hoch – Angebotskräfte jetzt stärken.

Joint Economic Forecast (2022), Klimaschutz ohne Produktionseinbußen: Die Rolle energiesparenden technischen Fortschritts, Hintergrundpapier zur Gemeinschaftsdiagnose Frühjahr 2022.

Joltreau, E. and K. Sommerfeld (2019), Why does emissions trading under the EU Emissions Trading System (ETS) not aﬀect ﬁrms’ competitiveness?, Empirical ﬁndings from the literature, Climate Policy Vol. 19, No 4, pp. 453‑471.

Jüppner M., A. Martin and L. Radke-Arden (2024), The Effects of Energy Efficiency on GDP and CO2 Emissions in Germany, Deutsche Bundesbank Technical Paper, forthcoming.

Känzig, D. R. (2023), The Unequal Economic Consequences of Carbon Pricing, National Bureau of Economic Research, Working Paper No 31221.

Karmaker, S. C., S. Hosan, A.J. Chapman and B.B. Saha (2021), The role of environmental taxes on technological innovation, Energy, Vol. 232, No 121052.

Kriegler, E., J. P. Weyant, G. J. Blanford, V. Krey, L. Clarke, J. Edmonds, A. Fawcett, G. Luderer, K. Riahi, R. Richels, S. K. Rose, M. Tavoni and D. P. van Vuuren (2014), The role of technology for achieving climate policy objectives: overview of the EMF 27 study on global technology and climate policy strategies, Climatic Change, Vol. 123, No 3‑4.

Metcalf, G. E. and J. H. Stock (2023), The Macroeconomic Impact of Europe’s Carbon Taxes, American Economic Journal: Macroeconomics, Vol. 15, No 3, pp. 265‑286.

Network for Greening the Financial System (2023a), NGFS Scenarios for central banks and supervisors.

Network for Greening the Financial System (2023b), NGFS Climate Scenarios, Technical Documentation.

Newell, R. G., A. B. Jaffe and R. N. Stavins (1999), The Induced Innovation Hypothesis and Energy-Saving Technological Change, Quarterly Journal of Economics, Vol. 114, No 3, pp. 941‑975.

Nordhaus, W. D. (2013), The Climate Casino: Risk, Uncertainty, and Economics for a Warming World.

Office of Energy Efficiency and Renewable Energy (2022), Energy Efficiency, Definition of Energy Efficiency by the Office of Energy Efficiency and Renewable Energy.

Popp, D. (2002), Induced Innovation and Energy Prices, American Economic Review, Vol. 92, No 1, pp. 160‑180.

Stern, D. I. (2019), Energy and economic growth, in: Routledge Handbook of Energy Economics, pp. 28‑46.

Footnotes

By 2030, emissions have to be reduced by 65 % compared with 1990 levels. They had already been cut by around 46 % by 2023.
In its “Fit for 55” package, the European Commission envisages a 55 % reduction in EU emissions by 2030 compared with 1990 levels. The stringency of targets in each country depends on the size of the nation’s economy.
See Office of Energy Efficiency and Renewable Energy (2022).
Still, there is disagreement in the empirical literature as to the aggregate impact of higher carbon prices. While Metcalf and Stock (2023) arrive at a low estimate for the aggregate impact of higher carbon prices, Känzig (2023) finds they have a dampening effect on economic activity.
See Acemoglu et al. (2012) and Network for Greening the Financial System (2023a).
Newell et al. (1999) and Popp (2002) show that higher energy prices increase the incentive to develop energy-saving technologies. See Acemoglu et al. (2012) for a structural analysis.
This could be reinforced by the government also promoting the development of technologies that contribute to reducing emissions alongside carbon pricing. See Acemoglu et al. (2012). This channel is also disregarded by assumption in the analysis.
Empirical studies indicate that carbon pricing has a positive incentive effect on innovation. See Karmaker et al. (2021) and Aghion et al. (2016). However, it is difficult to estimate the strength of additional energy efficiency gains via this channel.
The EMuSe model was developed at the Bundesbank and is a dynamic stochastic general equilibrium (DSGE) model augmented by a multi-sectoral production structure and an environmental module. See Hinterlang et al. (2023) for technical documentation on the model.
The two components are analysed separately in the model calculations. They can be combined in principle.
Studies for the United States show, for example, that changes in energy-saving technology are an important driver of developments in emissions (Nordhaus (2013) and Jo and Karnizova (2021)).
See Kriegler et al. (2014) or Bönke et al. (2023).
There are multiple viewpoints in the academic debate about how well energy can be saved using other factors of production within production processes. Essentially, though, energy plays a special role in the production process and its degree of substitutability is very limited, at least in the short term (Stern (2019)).
This approach is described in detail in Hassler et al. (2021).
Based on the NACE classification, the following sectors are defined: Agriculture (A), parts of the manufacturing sector not covered by the EU Emissions Trading System (EU ETS), EU ETS parts of the manufacturing sector excluding the manufacture of coke and refined petroleum products (C17, C20, C23, C24), water supply (E), construction (F), wholesale and retail trade (G), transportation and storage (H), other services (I-N, R, S), and two energy sectors. There is a fossil energy sector comprising mining and quarrying (B) as well as the manufacture of coke and refined petroleum products (C19), and there is the electricity and gas supply sector (D35).
This figure is consistent with the results presented in Joint Economic Forecast (2022) and Bönke et al. (2023).
Energy intensity is measured as primary energy consumption relative to GDP, in line with the COP28 definition.
Rebound effects may weaken the reduction in energy intensity compared with energy efficiency improvements, meaning that not all energy savings that would be technically possible are actually implemented. Once energy can be used more efficiently, production costs can be cut because energy consumption and the price of energy drop. These cost savings make it possible to increase production. As a result, demand for energy goes up again somewhat relative to other factors of production, and energy intensity falls somewhat less than it does immediately after the increase in energy efficiency.
See, for example, Bruns et al. (2021) and Jo and Karnizova (2021).
The importance of sectoral interlinkages for aggregate effects, particularly in the case of the introduction of a carbon price, was explained in Deutsche Bundesbank (2022).
See, for example, Network for Greening the Financial System (2023b).
This distinction is based on the model variant in Hinterlang et al. (2022).
The analysis excludes economic damage caused by an excessive concentration of greenhouse gases in the atmosphere. As greenhouse gas emissions from the rest of the world are not included in this analysis and greenhouse gas emissions in Germany (and any reductions thereof) represent only a fraction of total emissions, it can be assumed that the impact over the period under consideration is negligible. The analysis in Deutsche Bundesbank (2022) on the physical impact of climate change on the economy as a whole shows gradual global warming having a very small effect on the German economy up until 2020. The effects of other physical risks, such as extreme weather events, were not investigated.
The results presented here refer to Jüppner et al. (2024).
Aggregate effects stemming from the allocation of resources – for example, towards research and development of energy-saving technologies – are excluded. Hulten (2001) describes technological progress of such an exogenous nature as “manna from heaven”.
There are studies that analyse the impact of carbon pricing in the context of the European Union Emissions Trading System (EU ETS) on the overall technological progress (total factor productivity) of enterprises or sectors. In some cases, only small or insignificant effects are found (see, for example, D’Arcangelo et al. (2022) and Joltreau und Sommerfeld (2019)). Evidence on the impact of carbon prices on energy-saving technological progress has only been examined to a limited extent thus far. When working with simulation calculations, it can be taken into account by investigating scenarios with different assumptions regarding the evolution of energy efficiency. For example, this is the approach employed by the Network for Greening the Financial System (NGFS) (see Network for Greening the Financial System (2023b)).
How revenue from carbon pricing can be used for climate-friendly investment is also not examined here. For more information, see, for example, Andrés et al. (2024). We also refrain from analysing the implications of international linkages, such as the spillover effects of energy-saving technologies or carbon pricing in Germany and abroad.
The carbon price in the model initially leans on the predefined price pathway of the national emissions trading system (nEHS) up to 2026. The carbon price in the model is set so that the price of fossil energy in the first year after the carbon price is levied reflects the percentage increase in the price of fossil fuels as a result of the carbon price. The carbon price of €45 levied in Germany caused fuel prices to rise by 6.75 % relative to 2019 (Federal Ministry of Finance (2022) and ADAC (2024)). As the impact of the pandemic in particular distorted prices in 2020, 2019 was chosen as the pre-carbon price reference year. The carbon price in Germany was introduced back in 2021 at a low level. However, since the model simulations first apply from 2024 and the model in the 2023 reference scenario does not include a carbon price, the carbon price is introduced at a higher level in the model and the increase in the price of fossil fuels caused by the higher carbon price is taken into account from 2024. The carbon price continues to rise in the model until 2026 in line with the actual price pathway, increasing by a rate of 22 % and 18 % in 2025 and 2026, respectively. Since a separate European emissions trading system is to be introduced for nEHS sectors from 2027, the price pathway is based on allowance prices under the EU ETS. From 2027 onwards, the carbon price climbs at an annual rate of 7.8 % until 2030 to reach the projected value of around €88, which corresponds to the average price for emission allowances in 2023 (Federal Environment Agency (2023)). There are still minor restrictions when it comes to the price pathway. First, there are sectors that are covered by the EU ETS and that already pay the higher carbon price. The carbon price for these sectors is thus somewhat lower between 2024 and 2026. Second, the fossil energy sector is taxed in the simulations. This sector is composed of sectors C19 and B. While sector C19 has to pay a carbon price, sector B has not yet been included in either of the two emission trading schemes.
Other contributions from the literature that predict a decline in output after the introduction of a carbon price include, for example, Hinterlang et al. (2022) and Bönke et al. (2023).
The model simulations do not make any further assumptions about economic developments up to 2030. It should be stressed that this is not an economic projection. The simulations focus exclusively on the contribution of the two components: energy efficiency improvements and carbon pricing.
The carbon price has considerable advantages when it comes to putting it into practice. It is easy to implement and functions as a tax on emissions that makes fossil fuels more expensive. It thus acts as an incentive, prompting a reduction in demand for fossil energy or emissions directly. No preliminary investment or funding programmes are necessary and the instrument’s impact unfolds almost immediately after its introduction (see Brand et al. (2023)). In addition, it can be adjusted flexibly depending on how effective it is being. Ultimately, it is also likely to have a positive incentive effect on investment in lower-emission technologies.
In the model simulations, energy-saving technology only exists in the corporate sector. Energy efficiency with regard to households’ consumption is assumed not to change. It is far more difficult to determine. Thus, the figure for energy-saving technology as an aggregate tends to be at the lower end. In principle, efficiency gains in the household sector could also contribute to reducing aggregate energy demand and thus energy intensity.
Further innovations, such as neutral technological progress, are not included in the simulations. Such advancements can trigger an increase in energy demand and thus also push up demand for fossil energy. Emissions would increase. If firms did not become more efficient in terms of energy use in this case, they would cut fewer emissions than in the status quo scenario.
A similar study conducted by Bönke et al. (2023) also concludes that a relatively high carbon price is necessary to achieve the emission targets if energy-saving technological progress evolves in the same way as in the past.
See Federal Environment Agency (2024).