The real substitution effect of renewable electricity: An empirical analysis for Germany

Highlights

• Onshore and offshore wind provide larger marginal emission reductions that solar electricity in Germany.

• A significant fraction of renewable electricity is exported.

• Exports lead to emission reductions abroad.

• Market Stability Reserve turns a fraction into long-term abatement.

• Extent depends on year of abatement and market outcomes.

Abstract

Renewable electricity is the backbone for a net-zero carbon society. This paper estimates the national and international emissions effect of German renewable electricity in the years 2017–2019 using a Random Forest algorithm, finding negative but heterogeneous effects on emissions demand. A fraction of the estimated emission reductions translate into allowance cancellations in the EU ETS and thereby reduce overall long-term emissions.

Introduction

The electricity sector in the European Union is one of the main contributors to the union’s greenhouse gas emissions. Currently, there are two policy levels targeted at decarbonizing the electricity sector. First, the supranational EU ETS, which covers power plants and industrial sites, requires regulated entities to surrender an allowance for every ton of CO₂ they emit into the atmosphere. Second, as preferences for tackling climate change appear to be heterogeneous across the European Union (Marchiori et al., 2017, Cetkovic´ and Buzogány, 2019), some member states have implemented their own, additional policies, such as setting energy efficiency targets or subsidizing the generation of renewable electricity with the goal of achieving a climate-neutral electricity sector. This has contributed to renewables providing just over 50% of German electricity demand in the year 2020 (Fraunhofer, 2021), up from six percent in the year 2000 (BMWi, 2021). At the same time, about 17% of European emissions in the electricity sector were emitted from power plants located in Germany (DEHSt, 2020), making it a particularly interesting case to study.

The two policy layers overlap and there are interaction effects. Before the EU ETS was reformed in 2018 by implementing the Market Stability Reserve (MSR), unilateral policies such as renewable subsidies only shifted cumulative emissions in time and space but did not reduce them, because the supply of allowances was fixed. The MSR alters this picture by permanently deleting allowances depending on the amount of unused allowances in circulation (Perino, 2018). Unilateral policies can now reduce total emissions - but at least in principle, they can also increase them (Perino et al., 2020, Gerlagh et al., 2021).

In light of these issues, this paper answers the question of how effective renewable electricity in Germany is in abating greenhouse gas emissions in the EU. To that end, I construct a rich dataset of hourly plant-level emissions and cross-bordert rade. I estimate the amount of offset emissions in response to higher renewable production using a novel estimation technique based on Machine Learning algorithms. Together with a detailed analysis on emission leakage via cross-border flows, this gives deep insights into the mechanics of the German electricity system, including the climate impact of grid congestion. I merge these results with recent insights about the behavior of the MSR to estimate the long-term abatement effect of German renewable electricity.

Onshore and offshore wind are more effective in reducing emissions than electricity from solar panels. Redispatch measures have only a slight impact on this magnitude. Negative leakage effects (Baylis et al., 2014) within the EU are substantial, reinforcing the reduction in emissions demand. This result is in line with internal carbon leakage for demand side policies¹ from Perino et al., 2020. The estimates range up to 91% for solar, meaning a domestic emission reduction is almost mirrored by Germany’s neighboring countries. A fraction of these emission cuts translate to long-term abatement under the EU ETS as a result of additional allowance cancellations. This portion approaches one with increasing duration of MSR storage for electricity produced before 2018, but decreases for later abatement.

It is well established in previous literature that the emission impacts of renewable electricity differs between conventional technologies.² Generally, wind is more effective than solar in reducing emissions. I extend the literature by studying onshore and offshore wind separately. These two technologies are heterogeneous in terms of intermittency and generation peak within the day. Both technologies tend to reduce emissions by less than the system average in all considered settings.

The technical efficiency of power plants is found to be negatively affected by renewable supply as a result of increased cycling (Kaffine et al., 2013). I construct a detailed plant-specific emission series that adjusts the often used average emission rate when the turbine is running at part load.

Abrell et al. (2019) compare relatively isolated Spain and highly interconnected Germany to highlight the relevance of inter-zonal electricity trade on domestic emission reductions. A high capacity for trade allows for sizable leakage. I use detailed data on international electricity trade that allows for specific calculation of reduced emissions in neighboring countries and thereby leakage effects.

A complementary carbon price under the EU ETS and its effectiveness in reducing emissions is empirically investigated for the EU ETS by Gugler et al., (2021) and for Texas by Cullen and Mansur (2017). Aside from endogeneity issues, the aspect of overlapping policy and its impact on long-term emission abatement is overlooked³ in this strand of literature. I offer a new angle by analyzing the interplay between these two policy levels.