• No results found

Bachelor's Thesis Economics - 2022 Emanuil Petrov- 12350737

N/A
N/A
Protected

Academic year: 2023

Share "Bachelor's Thesis Economics - 2022 Emanuil Petrov- 12350737"

Copied!
36
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

Residential Electricity Consumption after the European Green Deal Policy Objectives, Smart Prosumers and Flexible Demand Response

Bachelor's Thesis Economics - 2022

Emanuil Petrov- 12350737 Supervisor

Christopher Graser

(2)

Statement of Originality

This document is written by Emanuil Petrov (12350737) who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document are original and that no sources other than those mentioned in the text and its references have been used in creating it.

UvA Economics and Business is responsible solely for the supervision of completion of the work and submission, not for the contents.

30.06.2022 Emanuil Petrov

(3)

Table of Contents:

INTRODUCTION: ... 5

SECTION 1: DETERMINANTS OF RESIDENTIAL ELECTRICITY CONSUMPTION ... 7

1.1HOUSEHOLD ENERGY POVERTY IN THE EU: ... 7

1.2POLICY REVIEW: ... 11

1.3RENEWABLES IN THE ENERGY SYSTEM: ... 13

1.4PROSUMERS LITERATURE REVIEW:... 15

1.4.1 PV adoption: ... 15

1.4.2 Smart prosumers and prosumers aggregation: ... 16

1.4.3 Market paradigms: ... 17

1.4.4 Prosumer demand response profiles: ... 18

SECTION 2: DEMAND RESPONSE UNDER DYNAMIC COST OF CONSUMPTION ... 20

2.1TIME OF USE TARIFFS: ... 20

2.2LOW CARBON LONDON: ... 21

2.3RELEVANT DATA: ... 23

2.4METHODOLOGY: ... 23

2.5RESULTS AND DISCUSSION: ... 25

2.5.1 Constraint Management tariff: ... 25

2.5.2 Supply Following tariff: ... 25

2.5.3 Discussion: ... 26

CONCLUSION: ... 27

BIBLIOGRAPHY: ... 29

APPENDIX: ... 35

(4)

Abbreviation Term

CM Constraint Management

DER Distributed Energy Resources

dToU Dynamic Time-of-Use

EC European Commission

EP European Parliament

EU European Union

EV Electric-Vehicle

IHD In-Home-Display

IoT Internet-of-Things

IPP Independent Power Producers

LCL Low Carbon London

MO Market Operator

MS Member State

NGEU NextGenEU

P2P Peer-to-Peer

PPS Purchase Power Standard

PV Photovoltaic

RPEU RePowerEU

RRF Recovery and Resilience Facility

SF Supply Following

ToU Time-of-Use

TSO Transmission System Operator

V2G Vehicle-to-Grid

VPP Virtual Power Plant

(5)

Introduction:

Climate change and environmental degradation present an existential threat to Europe and the world. The European Commission (EC) and the European Parliament (EP) outlined tackling climate change and environmental-related challenges as "this generations defining task" (EC, 2019a). On December 11th 2019, in an effort to protect, conserve and enhance the European Union's (EU) natural capital, and protect the health and well-being of citizens from environment-related risks and impacts, the EC announced "The European Green Deal" – a roadmap for making the EU's economy sustainable by

"turning climate and environmental challenges into opportunities across all policy areas and making the transition just and inclusive for all" (EC, 2019b). All 27 EU Member States (MS) committed to reducing net emissions by at least 55% by 2030, compared to 1990 levels, and turning Europe into the first climate-neutral continent by 2050, decoupling economic growth from resource use and "leaving no one behind" (EC, NDa) (EC, NDb).

To reach its targets, Europe's decarbonisation efforts must run through all segments of the economy. Of particular relevance is the energy sector. The production and consumption of energy within the EU account for 75 per cent of total greenhouse gas emissions (EC, NDc). Since 2010,

electricity generation has accounted for around a fifth of total emissions. In the second quarter of 2021, the electricity supply accounted for 19 per cent of the total 867 million tons of CO2-equivalent

greenhouse gas emissions, a year-over-year sectoral increase of 17 per cent (Eurostat, 2021). The clean energy transition necessitates shifting away from a system based on fossil fuels to one dominated by non-polluting, renewable electricity.

Although fundamental progress has already been made in transforming Europe's electricity system, the production share of electricity generated by combustible sources is still high, standing at 41.3 per cent as of 2020 (Eurostat, 2022). What's more, as of 2018, Europe's energy import dependence, specifically of oil and gas, was at 55 per cent (EC, 2018). This heavy reliance on outside energy resources was one of the main factors behind the post-pandemic electricity price crisis. With the passing of the Omicron variant and the easing of lockdowns around the world, demand for goods and services began normalising. As a result, global demand for energy goods began rising. However, supply was unable to follow. After record floods struck the Chinese mining region of Henan in July 2021, Shanxi, the country's biggest coal-producing province, was also hit by flooding in October (BBC, 2021). Similarly, severe flooding impacted mines and key logistics routes in June in India's eastern and central states (The Economic Times, 2021). During the same period, OPEC increased oil production only marginally, not meeting expected future demand. Russian gas exports towards the EU fell by 24 per cent year-on-year (EC, 2022a). This led to a global mismatch between the supply and demand of energy goods, with prices for oil, natural gas, and coal, as of October 21st, rising by 63, 101, and 179 per cent, respectively (IG, 2021). The fourth quarter of 2021 also brought electricity consumption in Europe back to pre-pandemic levels, driven by a steady economic recovery (KPMG, 2022). The combination of these events drove electricity prices to unprecedented highs, with the European Power Benchmark averaging 194 €/MWh in Q4 2021 – 400% higher than Q4 2020 and 85% more than Q3 2021 (EC, 2022a).

After the liberalisation of the EU electricity market, regulated prices for the European residential sector in the majority of MSs were also lifted (Triconomics, 2020). Given the lack of protection and the skyrocketing prices, households faced overwhelming increases in their energy bills (Market Observatory for Energy, 2022). With energy poverty, defined as the inability to afford sufficient thermal comfort,

(6)

already affecting over 50 million people within the Union, this unprecedented volatility threatened to increase that number exponentially (EC, NDd). Although the transition toward renewables significantly improves the security of supply (EC, 2018), the new market realities showcased the fragility of the current EU energy system and the exposure of households to it.

Providing societies with reliable electricity distribution while mitigating energy poverty and climate change entails a crucial infrastructure component. Both the decarbonisation and the energy access issues require more decentralised energy solutions and a change in the energy infrastructure paradigm (Goldthau, 2014). To increase the resilience of the energy grid and promote self-consumption and flexible demand response (DR), on May 18th 2022, the EC announced its Solar Energy Strategy (EC, 2022b). Part of the REpowerEU (RPEU) initiative, it represents a building block, part of a complex policy cycle which began with the European Green Deal. With the Solar Energy Strategy, the EC proposes a mandate which will require all new residential buildings built after 2029 to be equipped with

photovoltaic (PV) units, thus transforming future households from idle consumers to active prosumers.

As prosumers are agents that both consume and produce energy, prosuming offers the potential for consumers and electric vehicle (EV) owners to re-evaluate their energy practices (Parag, 2016).

However, because physical energy infrastructure exhibits strong local, provincial, national and even supranational dimensions, alterations to it bring many unknowns and risks that need to be identified and managed (Goldthau, 2014). This implicit uncertainty, coupled with the rising risks of energy poverty, poses a challenge to policymakers. As such, the appropriate policy response must consider the near- future development trajectories of the critical determinants of households' electricity practices, namely the state of the electricity market, variable distributed energy resource (DER) integration, the clean energy transformation policy framework, and prosumer behaviour and aggregation.

Quantifying the mandate's effectiveness in promoting flexible consumption and mitigating residents' exposure to energy poverty requires a holistic approach to approximating the position of European households between the different dimensions of influence. What's more, such knowledge is fundamental to policymakers and planners, as integrating prosumers effectively and efficiently into competitive markets requires a broad spectrum analysis (Parag, 2016). Therefore, in an effort to provide clarity and inform policymakers, this paper examines the intricacies of this issue by separating the underlying determinants into two categories.

In Section 1, an extensive profile of the near-future EU prosumer household is developed based on current market trends, EC's policy initiatives and objectives, and the technical aspects of prosuming.

To those ends, a thorough analysis of the exposure of the residential sector to energy poverty will be carried out. Second, a classic-modernist governance paradigm and a policy cycle model will be employed to extract and underline the key elements of the EU's strategy for the clean energy transition and the mitigation of households' exposure to the volatility of electricity prices. Next, system-level analysis of the benefits and challenges of distributed variable renewable energy integration will be conducted. To conclude, a literature review on prosumers behaviour, smart prosumer aggregation, market paradigms, and demand response will be presented.

After establishing the characteristics framework of the residential sector, in Section 2, a quantitative analysis will be carried out to evaluate the efficacy of the EC's PV mandate on enabling flexible DR. To those ends, this paper contributes to the academic literature by transposing the normal and smart prosumer demand response incentive structures onto normal consumers, by drawing

(7)

parallels between them and time of use (ToU) tariffs. Next, data from the most extensive system

resilience-orientated, dynamic ToU trial in the UK will be used to quantify households' responsiveness to flexible demand response incentives. After formulating the control and treatment groups, season and time of day variables will be defined, based on which several mean consumption two-sample t-tests will be carried out to examine consumer responsiveness under different conditions. Finally, in the discussion sub-section, by combining the econometric results and the holistic framework from Section 1, a

discussion on the future of European households following the clean energy transition will be carried out. The benefits and limitations of the employed methodology will also be outlined. Additionally, the findings will be compared to the results of real-life studies.

Section 1: Determinants of residential electricity consumption

1.1 Household energy poverty in the EU:

Establishing the current state and trends in the European electricity market and their respective relation with household consumption is a natural foundation for the development framework, as market failure predisposes policy developments and residential behaviour.

The share of electricity in final energy consumption has increased by 5 per cent over the last thirty years (EEA, 2019). Despite this historically marginal increase, electricity demand is projected to grow significantly on a pathway toward climate neutrality (JRC, 2020). Its share is forecasted to increase from 23% today to around 30% in 2030 and towards 50% by 2050. In the residential sector, primarily due to the policy-induced roll-out of electric heat pumps, the share of electricity in heating demand is expected to increase to 40% by 2030 and 50-70% by 2050 (JRC, 2020). The substantial growth in household electricity consumption risks exposing a broad economic demographic to potential energy poverty. Energy poverty is already a widespread problem across Europe, affecting between 50 and 125 million people (EC, NDd). The inability to afford sufficient thermal comfort, in turn, translates into a worsening socio-economic situation, severe health issues and social isolation (Sánchez et al., 2016).

Nominal retail electricity prices for households in the EU27 have increased by a third from 156 EUR/MWh to 208 EUR/MWh from 2008 to 2019 (Triconomics, 2020). Since 2015 the average prices for European households have been the highest of all G20 countries. Price developments in all MS have been adverse for end consumers. Although this may not affect industry competitiveness within the Union, it represents a worsening of the relative price paid by the average household (Triconomics, 2020). In 2018, European households in the lowest 10 per cent income bracket spent 8.3 per cent of their expenditure on energy. Lower-middle and middle-income households spent 7.4 and 6.7 per cent, respectively (Triconomics, 2020). While northern and western European middle-income households spent between 3 and 8 per cent, the expenditure for the same income level of their central and eastern European counterparts is between 10 and 15. The poorest households in Slovakia and Czechia spent around 20 per cent, while their counterparts in Luxembourg, Finland and Sweden spent less than 5 per cent (Triconomics, 2020). However, it is essential to consider that purchasing power varies between MS.

(8)

Figure 1. (Triconomics, 2020). Household electricity prices, EU27, other G20, 2008-2019, EUR2018 /MWh

Figure 2. (Triconomics, 2020). Comparison of EU27 weighted average with G20 (trade) weighted average.

Analysing electricity prices in terms of Purchasing Power Standards (PPS) gives a better measure of the relative effect of prices on households' finances. As prices are adjusted in an artificial common reference unit under PPS, it allows for comparison based on purchasing the same volume of goods and services, thus removing the price level effect (Insee, 2021). As a result, the relative differences present a clearer understanding of the actual impact on households in each country (Triconomics, 2020).

(9)

Figure 3. (Triconomics, 2020). Comparison of 2017 retail household electricity prices, nominal and PPS, EUR/MWh

In PPS adjusted terms, the unit price borne by households in half of the MS is higher than in nominal terms, while for the other half is lower. On average, high-income Eurozone states see their ranking decrease, meaning electricity is cheaper. Inversely, countries with income levels lower than the EU average see their prices increase. Post-soviet states, in particular, see the highest relative increases.

The trend of retail electricity prices in the EU for household consumers worsened following the recovery from the Covid-19 pandemic. Since July 2021, price divergence began increasing, surging in the fourth quarter (MOE, 2022). In fact, household prices reached the highest level of divergence on record in all consumption bands (EC, 2020a). The Market Observatory for Energy concluded that the main factor behind the increase in 26 out of 27 EU capitals was the rise in wholesale prices. As inflation is strongly correlated with wholesale electricity prices, household prices are not expected to plateau until the global inflationary pressures subside (Australian Energy Council, 2022).

(10)

Figure 4. (MOE, 2022). The standard deviation of retail electricity prices in the EU for household consumers.

Figure 5. (MOE, 2022). The Household Energy Price Index (HEPI) in European capital cities in Eurocents per kWh, February 2022.

With retail electricity prices for households already being the highest of all G20 countries, the standard deviation reaching historic levels, and the forecasted growth of the share of electricity in final energy consumption, the future of European residents looks bleak. Energy poverty, already a significant problem within the Union, threatens to escalate further, disproportionally affecting the economically weaker MS. Understanding the severity of this issue requires understanding the policy environment in which it will develop. As households' exposure to the market is primarily determined by legislation,

(11)

residential welfare is dependent on the EU's energy policy developments and the Commission's response to the electricity crisis.

1.2 Policy review:

Melanie Mitchell (2009) defines a complex system as "a system in which large networks of components with no central control and simple rules of operation give rise to complex collective behaviour, sophisticated information processing, and adaptation". The models of complex adaptive systems, such as the electricity market, implicitly assume that the system is a closed one and as such, regulation is possible as long as patterns within the system can be identified (Oğuz,2020). The EC's ability to impose market and non-market (prescriptive) regulations allows for the alteration of the "rules of operation", thus resolving welfare losses. To obtain the envisioned rule structure, or policy objective, identifying the underlying pattern which the inefficiency arises from, referred to as problem definition, is critical. As such, analysing the EU's energy policy developments is fundamental in establishing the functional form specification of the electricity market in which future prosumers will participate.

Achieving climate neutrality necessitates the decarbonisation of all sectors of the economy. This includes agriculture, manufacturing & construction, transport services, household heating and electricity generation. To succeed, the EU needs a cohesive and detailed policy design related to the resolution of these issues. Mara S. Sidney (2007) defines policy design as "identifying and/or crafting a set of policy alternatives to address a problem and narrowing that set of solutions in preparation for the final policy decision". Public policy theory, in particular the classical-modernist governance paradigm, perceives policy design as a rotation between several stages: agenda setting; problem definition; policy alternatives; policy implementation; re-evaluation based on results achieved, updated problem definition, etc. (Knill & Tosun, 2020). The future of European households is dependent not only on the ongoing trends in the electricity markets but also on the policy environment and government

interventions. As such, understanding its trajectory requires identifying the key elements in the policy design cycle.

The clean energy transition found its regulatory beginnings in the European Parliament with the introduction of Regulation (EU) No 2019/943 of June 5th, 2019, on the internal electricity market. The regulation establishes the fundamental normative principles for an efficient electricity market fit for the clean energy transition, namely "to deliver real choice for all Union final customers, be they citizens or businesses, competitive prices, efficient investment signals and higher standards of service, and to contribute to the security of supply and sustainability." (EC, 2019c). With this regulation, the EC

decisively set the agenda for future policy cycle developments. Paragraphs 5 and 7 outline the necessity for "small power-generating facilities [to] be granted priority" and to transform "electricity customers [which] were purely passive ... in the future need to be enabled to fully participate in the market on equal footing with other market participants", respectively.

On December 11th 2019, all 27 EU Member States (MS) committed to reducing net emissions by at least 55% by 2030, compared to 1990 levels, and turning Europe into the first climate-neutral

continent by 2050, decoupling economic growth from resource use and "leaving no one behind" (EC, NDa) (EC, NDb). With this agreement, the EP and the Council of Ministers defined the problem which will be tackled in the later stages of the policy cycle. The European Green Deal, the blueprint for this

(12)

transformational change, outlined a number of policy alternatives to reach the Union's goals. For the clean energy transition, the production and consumption of which account for more than 75% of the EU's greenhouse gas emissions, it focuses on three fundamental principles (EC, NDc). These principles are: ensuring a secure and affordable energy supply; developing a fully integrated, interconnected and digitalised energy market; and prioritising energy efficiency and a power sector based largely on renewable sources (EC, NDc).

Not long after Regulation (EU) No 2019/943 and the European Green Deal were ratified by the EP, the Covid-19 pandemic began. The pandemic challenged and altered the fundamental mechanism through which the European economies functioned, interacted with each other and conducted business.

The sudden change in everyday practices and behaviour of economic agents on all levels required policymakers to re-evaluate the policy alternatives through which to attain the objectives outlined in the European Green Deal. After Italy's initial lockdown, on March 9th 2020, the Members of the European Council adopted a joint statement calling on the European Commission to "develop a coordinated exit strategy, a comprehensive recovery plan and unprecedented investment to allow a normal functioning of our societies and economies and get to sustainable growth, integrating inter alia the green transition and the digital transformation" (EC, 2020b). To protect its citizens and overcome the economic

challenges, the Union needed a revised growth strategy which promotes modern, resource-efficient and competitive markets decoupled from carbon emissions.

The EC, in collaboration with the European Council of Ministers, recognised not just the severity of the moment but also the opportunity to emerge stronger from the pandemic and make Europe greener, more digital and more resilient. Motivated by a common goal, on July 21st, 2020, MS agreed upon the temporary recovery instrument NextGenerationEU (NGEU). Worth more than €800 billion, at current prices, NGEU provided the necessary funding not only to deal with the short-term consequences of the pandemic but also to set the EU on a trajectory in line with the European Green Deal (EC, NDe).

The main component of the NGEU is the Recovery and Resilience Facility (RRF). The RRF's goal is to allocate funds to MSs to assist with implementing reforms and investments that align with the EU's policy priorities (EC, NDf). Amounting to €723.8 billion in loans (€385.8 billion) and grants (€338 billion), MSs have allocated almost 40 per cent to climate and energy-related measures (EC, NDf).

In 2022, following Russian President Vladimir Putin's unprecedented invasion of Ukraine, the electricity market experienced even higher volatility. Furthermore, uncertainty in the market came to historic highs after Gazprom unilaterally stopped gas deliveries to Poland, Bulgaria, the Netherlands, Denmark, Finland, and Germany, something that did not transpire even in the Cold War (The Guardian, 2022). On March 8th, in an official communication to the EP, the EC recognised the significant harm being done to consumers, namely "that the retail market is failing to protect consumers" in a way "likely to harm longer term market developments and undermine the energy transition" (EC, 2022c). The new geopolitical and market realities underlined the necessity for the re-evaluation of the different elements of Europe's clean energy transition and energy independence policy approach once more. Building on the legislative framework of the European Green Deal and the liquidity provided by NGEU and the RRP, on May 18th, the EC announced the REPowerEU initiative (RPEU) (EC, NDf). The plan aims at mitigating the macro-economic and social impacts of high energy prices, which risk dragging households into energy poverty, by reducing the EU's dependence on Russian fossil fuels and increasing the resilience of the EU's energy system. In line with these objectives, the EC has proposed to increase the binding

(13)

threshold of renewable sources in the EU's energy mix to 45% by 2030 (EC, NDa). Additionally, the EU is working with international partners to find alternative energy supplies in the short term.

Regarding households, with the RPEU, the EC focuses on accelerating the clean energy transition and empowering consumers, by setting "measures that enable self-consumption and production" and

"tapping into the potential of demand-side flexibility" as their policy objectives, thus further

strengthening resilience against future shocks (EC, 2022a)(EC, 2022c). Regarding household electricity bills, the EC outlines the reduction of energy consumption as essential to bring down both emissions and energy costs for consumers (EC, 2022c). Furthermore, the EU-wide European Solar Rooftops Initiative, announced in the Commission's REPowerEU Communication, adopts provisions to ensure that all new buildings are "solar ready" and make the installation of rooftop solar energy compulsory for all new public and commercial buildings with useful floor area larger than 250 m2 by 2026; all existing public and commercial buildings with useful floor area larger than 250 m2 by 2027; all new residential buildings by 2029 (EC, 2022b).

The EC's decision to use prescriptive, non-market regulation guarantees market compliance and targeted effort towards the outlined policy objectives. Furthermore, as the PV technology is widely available and installation standards are analogous, wide scale adoption of the practice is feasible, relatively cheap and does not require significant alterations to building practices (Varela, 2022).

However, the integration of potentially millions of DER into the existing energy grid possess numerous challenges and requires a fundamental change in the way it operates.

1.3 Renewables in the energy system:

The recent decline in the cost of renewable energy technologies, the digitalisation of the economy and emerging technologies in batteries, heat pumps, electric vehicles, and hydrogen offer an opportunity for a profound transformation of the energy system and its structure (EC, 2020c). Between 2010 and 2020, the relative share of renewable electricity generation in relation to Unions net

production mix grew from 19.8 % to 34.1 %. The proportion of net electricity generated from PV's and wind turbines increased significantly, from 0.8 % to 5.3 % for solar and from 4.9 % to 14.7 % for wind power. In the same timeframe, the production share of electricity generated by nuclear power plants decreased slightly from 28.7% to 24.3%, while the share of combustible fuels saw a relatively significant decrease from 51.3% to 41.3% (Eurostat, 2022). These shares vary greatly between the different MS.

Within the EU, each MS determines its own energy mix in respect of the rules of the internal energy market (EC, NDf).

Structurally, each state has one or more internal electricity pools, e.g. DK1 & DK2, which serve eastern and western Denmark, acting as a centralised form of market management. In them,

Independent Power Producers (IPP), from national utilities to prosumers, place generation bids and retailers place consumption offers at the same time. Consumption offers are ranked in decreasing order, while supply offers are ranked in increasing order based on price. Afterwards, the Market Operator (MO) constructs supply and demand curves based on a merit order and utilises a market-clearing algorithm that decides the scheduled bids and offers. All IPPs get the same price for the electricity they are generating at the given time block. This type of wholesale market is referred to as a system of marginal pricing or pay-as-clear. As renewable energy sources by definition have a marginal cost of 0,

(14)

they are grid parous and always scheduled. Despite the implied volatility of renewable power

generation, clearing a market with renewable energy offers gives rise to two fundamental properties.

Firstly, the cleared energy volume is at least as much as in the case of no renewable energy. Second, the clearing price is, at worst, the same as in the case of no renewable energy (Pinson, 2018).

As the supply and demand of electricity within the grid always must be matching, Transmission System Operators (TSO) are responsible for ensuring the power balance of the system. Off-balance may arise due to: electric load or renewable energy generation being greater or lesser than foreseen at the time of market-clearing, outages, or internal congestion within the market/balancing zone (CRU, 2019).

As such, balancing is a close to real-time operation necessary to ensure power system quality and stability. To those ends, TSOs make use of ancillary services. Ancillary services are any type of service that supports power system operations directly bought by the system operator, e.g. Primary reserves, Secondary reserves, Tertiary reserves, and voltage control (Pinson, 2018). The quantity of ancillary services needed is determined based on the uncertainty in electricity demand, uncertainty in renewable energy generation, and operational constraints, e.g. commitment and ramping capabilities. As such, managing a power system with significant renewable electricity production within its respective energy mix requires TSOs to contract high amounts of ancillary services, hence lowering available market supply and increasing the end price on the wholesale market. As the proportion of variable renewable energy generation increases, so does the implied uncertainty, thus necessitating higher reserved capacity.

The dogmatic relationship between uncertainty and needed capacity is unavoidable in isolated markets. However, the EU's electricity market follows an unconventional integrated structure, which found its legislative beginnings in the early 1990s. Significant benefits were achieved thanks to the gradual market integration, which culminated with the Clean Energy Package (Roques, 2020). The interconnectivity between different internal markets introduces more competition and decreases the absolute levels of additional capacity needed (Bockers et al., 2013). Welfare efficiency gains are then realised due to the decreased concentration of the wholesale market and the reduction of necessary idle generation capacity. The divergence of high-peak periods between MSs reduces the need for contracting ancillary services as they can be traded between markets. The higher the divergence, the higher the utility gained from capacity reserves. However, as renewable energy generation within the mix increases, market integration must follow, or else the net welfare gains will diminish or be lost altogether. As such, the flexibility needed for the overall management of an energy system envisioned by the European Green Deal necessitates further integration of the electricity market.

Europe's energy future relies on an ever-growing share of geographically distributed renewable electricity generators and flexibly integrated energy carriers. The coordination and planning of

operations in the electricity system as a whole is fundamental in realising the goal of affordable and decarbonised electricity supply. The RRF not only provided funding for investment in clean technologies and value chains but also highlighted the need for better integration of the electricity system (EC, 2020d). Traditionally, the electricity system is based on a unidirectional market mechanism (Oğuz, 2020). It was built this way as historically, electricity is produced at one end, in a small number of large generation facilities, and transmitted through the existing grid to the other end, namely to end

consumers. However, modern systems must support multidirectional integration, in which consumers play an active role in the electricity supply. Vertical integration relates to decentralised production units and prosumers actively contributing to the overall balance and flexibility of the system by supplying excess electricity back into the grid on a local level or by downwards regulating via "vehicle-to-grid"

(15)

services (Simshauser et al., 2015). Horizontal integration, on the other hand, refers to exchanges of energy between the consuming sector, such as prosumers part of energy communities trading between each other (Quartierstorm, 2020). By linking up different electricity carriers, localised production, self- production and smart use of distributed electricity, system integration can contribute the greater consumer empowerment, improved resilience and security of supply (EC, 2020c).

1.4 Prosumers literature review:

Prosumers, i.e. consumers that also act as producers of energy, are seen as vital for the acceleration of the renewable energy transition (Kotilainen et al., 2016). The possibility of onsite production and flexible consumption transforms agents from passive consumers to active service providers in power systems (Grzanic, 2022). However, due to the complex nature of prosuming practices, policymakers seeking to employ them as part of their energy poverty prevention strategies require knowledge of how prosumers could be integrated effectively and efficiently into competitive electricity markets. To those ends, this paper conducts a literature review on the intrinsic and extrinsic motivations of prosumer households, smart prosumer aggregation and subsequent market integration, and demand response profiles.

1.4.1 PV adoption:

Households, businesses, communities, municipalities and other agents with PV installations rely on smart meters for the monitoring and management of their generation and outside supply of

electricity. Further, they may incorporate home energy management systems, energy storage, electric vehicles and vehicle-to-grid (V2G) systems with their installation to evoke a synergistic effect and increase their respective efficiency of consumption (Parag, 2016). As such, prosuming incentivises consumers to re-evaluate their energy practices. Given the prospect of substantial savings on their electricity bills, rational utility maximising agents will optimise their electricity use and match it with their electricity generation and storage capabilities, when applicable. Economic gains have been found as the main driver behind private investment in PV in Austria (Fleiß et al., 2016). In turn, the ability to capture these gains and their respective magnitude depends on household resources and market conditions, e.g., retail prices and subsidy schemes. Both higher income and higher costs of substitution have been found to be highly correlated with adoption rates (Hansen, 2022).

Beyond financial incentives, motivational and contextual factors affect adoption rates

significantly. Based on Danish survey data from a pool of over 75 000 households, Hansen (2022) finds that 40.3 per cent of adoptees are technically educated. Additionally, 62 per cent of survey respondents listed being a "technical frontrunner" as an important factor behind their investment decision. In the context of the Danish market, the management of social relations also plays a crucial role, expressed in the form of the so-called peer-effect. PV adoption can be interpreted as a signal of personal values to others, with 36 per cent listing "to set a positive example to others" as important to them.

(16)

Pro-environmental values are also highly correlated with PV uptake (Best, 2019). Briguglio (2017), by utilising the concept of "warm glow" ultraism as a theoretical construct, explains the rationale behind households' decisions to contribute to the environment instead of free-riding on the efforts of others. Wittenberg and Matthies (2016) find environmental concerns as predictors of engaging in load shifting and consumption reduction in German PV households.

Based on their different intrinsic and extrinsic motivations, such as financial gains,

environmental concerns, and self-sufficiency, prosumers will employ their respective generation and consumption capabilities in different ways. When aggregated, this gives rise to prosumer groups with fundamentally different structural and goal-oriented functional forms.

1.4.2 Smart prosumers and prosumers aggregation:

Smart management refers to the amalgamation of the different Internet-of-things (IoT) components which allow prosumers to actively and deliberately manage their respective consumption and production. These include in-home load displays, smart thermostats and HVAC systems, self- regulating batteries, smart wet appliances, smart electric heating pumps, etc. These smart elements predispose an important divergence between conventional grid consumers and smart prosumers in a number of dimensions. Firstly, when considering the aspect of resilience, following energy system voltage faults and outages, conventional consumers may only respond reactively in an effort to protect their assets and limit damage. Smart prosumers, however, can automatically detect and respond to distribution problems, thereby focusing on prevention (Parag, 2016). Additionally, while traditional consumers are uninformed and non-participative in the power system, prosumers are informed and actively involved (Hansen, 2022). Lastly, regarding diversification, consumers rely on large centralised IPPs and the central grid for their supply with little opportunity for self-storage. Smart prosumers, on the other hand, not only may rely on their own production and batteries but are also incentivised to

interconnect with each other, thus reaping the full benefits of distributed generation and storage (Lovaas, 2009).

The interconnection between different smart prosumers gives rise to autonomous, decentralised networks. They may be utilised for community or profit-maximising purposes. The prosumer-centric paradigm brings diverse possibilities for community-based prosumer groups and organisations. By forming microgrids, they can manage their needs in an efficient manner, given the dynamic changes in stakeholder needs, capabilities of local balancing resources, and prosumer services available (Karnouskos, 2009). A substantial prosumer base supports energy efficiency in smart grids by enabling demand response (DR) and flexibility, as such, the success and sustainability of the grid are highly dependent on the magnitude of prosumer participation. (Kotilainen et al., 2016). A flagship program by the Swiss Federal Office involving 37 participating households and a retirement home with PV installations in the town of Walenstadt found that through the establishment of a peer-to-peer trading microgrid, local energy trading almost doubled the community's self-sufficiency rate (from 21%

to 39%) and self-consumption rate (from 34% to 62%) (Quartierstrom, 2020). Such forms of local trading reduce transmission costs and grid loss while generating efficiency gains and redistributing profits fairly within the community (Metz, 2007) (Li, 2020). Additionally, the reduced exposure to the centralised electricity grid enhances the community's resilience against power outages (Goldthau, 2014).

(17)

Profit-maximising smart prosumers may choose to aggregate their distributed energy resources within Virtual Power Plants (VPP) and compete in the wider electricity market. Agents within a VPP include not just different types of distributed generation units but also active consumers and storage technologies connected via smart grids. The green energy transition involves the transformation of the generation portfolio, peak production, grid infrastructure, and consumption patterns. As such, by bundling flexible demand and supply elements together, VPPs can operate in a sustainable, profit- maximising and stabilising market-oriented mode (Loßner, 2017). Unlike microgrids, different agents need not be located in the near vicinity of one another, as long as they have proper smart components (Sučić, 2011). The main function of the VPP involves the scheduling and trading optimisation of its portfolio. The optimal trading strategy involves not only scheduling its DER based on numerical weather forecasts but also using its storage capital for upwards and downwards regulation. Additionally,

economic value can be extracted by selling renewable energy at a premium to interested consumers via over-the-counter exchange. Scenario-based forecasting, given the Mixed Integer Cost Optimization of Energy Systems for Europe model, MICOES-Europe, concludes that VPP revenues are between 11 and 30 per cent higher for cooperating agents as compared to their individual market participation (Loßner, 2017).

1.4.3 Market paradigms:

In light of the clean energy transition and the emerging prosumer phenomenon, Parag (2016) presents two pathways low-carbon systems may develop in regard to their connection to the wider grid.

The first path involves self-sufficient agents managing their electricity production and consumption detached from the grid. Given this paradigm, the market structure may follow either disconnected smart microgrids, referred to as "island mode", or a peer-to-peer (P2P) trading marketplace. P2P markets involve decentralised, flexible networks which emerge and are managed from the bottom up. P2P platforms would allow prosumers to place production, consumption, and ancillary service bids much in the same way as in electricity pools. Following the scheduling issued by the system, the distribution grid is paid a fee for its management and distribution functions. The decentralised nature of such a market structure, however, gives rise to the issue of accountability. While in the current electricity system, TSOs and DSOs carry the liability for outages and voltage spikes, the question of who is accountable for the safety and availability of electricity services in decentralised markets poses a great challenge (Parag, 2016).

(18)

Figure 6. (Parag, 2016). (A) Peer-to-peer market; (B) Prosumer-to-grid microgrids; (C) Prosumer-to- 'islanded' microgrids; (D) Virtual Power Plants.

Note: A dot represents a prosuming agent; A interconnection represents a transaction of prosuming service; A circle represents an organised group of prosumers.

The second pathway involves the integration of prosumers in the current grid and, as such, transforms them from passive consumers to active providers of energy services. Following this

paradigm, prosumers may pool their resources in the form of a VPP as discussed above or provide their services to a microgrid, which is in turn connected to the main grid. Microgrid functioning need not be community orientated. Corn et al. (2014) have developed a market-orientated microgrid operational system based on willing-to-participate users. In it, prosumers may choose to submit offers, given their willingness to adapt their respective loads, in return for monetary benefits. Afterwards, based on market dynamics, the microgrid operator may choose to accept or reject their offer. The main difference

between market-oriented microgrids and VPPs is the commitment on the part of the prosumer. While VPP agents must relinquish full control over their production and consumption units to the network operator for the duration of their contract, in microgrids, agents may choose when to offer their services.

Irrelevant of the path they choose, presuming enables agents to save on energy bills while contributing to wider social welfare (Parag, 2016).

1.4.4 Prosumer demand response profiles:

Demand response is seen as one of the most cost-effective flexibility sources in the electricity system, critical in enabling the integration of a high share of variable renewable generation (Enefirst, 2019). As discussed above, the willingness of consumers to shift demand is in its largest part related to

(19)

financial incentives. The final price paid by prosumers for their use of electricity equals the quantity of electricity supplied by their energy provider times their respective contractual rate per kWh. As the marginal cost for consuming any self-production equals 0, households with PV installations have a strong incentive to shift their loads to periods of time with high individual generation. However, as smart prosumers may choose to participate in smart grids, which provide DR incentives based not only on current solar irradiation but also on several other dynamic factors, an important distinction can be made between the DR profiles of households with PV installations and smart prosumers.

Figure 7. (Pimm, 2018). Average UK household Figure 8. (Rehman, 2017). PV generation profile.

electricity consumption.

Figure 7 shows the average electricity demand of UK households against the time of day for weekdays and weekends in Summer and Winter. Juxtaposed against the average PV generation profile, as displayed in Figure 8, a clear divergence can be seen between demand and supply. As solar

proliferation patterns are close to identical throughout the year, households with PV installations are incentivised to adapt their consumption behaviour in line with generation. However, following the initial alignment, load spreads for those households may remain static. As the mean production patterns don't change, neither should mean consumption.

However, to attain the maximum benefit from their PV installations, integrated smart

prosumers need to adapt their behaviour not only in accordance with solar generation patterns but also with wider market events related to network balancing. As such, their respective incentive structure for DR is dependent on the dynamic needs of stakeholders.

As outlined in the policy review, the EC's RPEU policy objectives relate to "measures that enable self-consumption and production" and "tapping into the potential of demand-side flexibility". While the PV mandate ensures the accomplishment of the first objective, the presence of solar panels alone does not incentivise a "flexible" demand response. It can be argued that the market will step in and create the environment for smart prosumers to interconnect and engage in flexible DR. However, even if smart prosumers are given this opportunity, the question of whether they'll engage in it arises. One way to study this behaviour is through the use of dynamic time-of-use tariffs.

(20)

Section 2: Demand response under dynamic cost of consumption

2.1 Time of use tariffs:

Currently, most households in Europe are supplied electricity by their respective energy provider based on flat tariffs (ACER, 2021). This means that households pay the same amount per kWh regardless of what time of the day or year they consume it. Given these types of contracts, the utility function of the residential sector exhibits a linear cost constraint. Time of use (ToU) tariffs, on the other hand, adjust the cost of electricity in a given time block based on wider market conditions. As such, by adjusting the price, energy providers change the demand. Given a set of market conditions, ToU tariff programmes can be used to shift demand towards periods when renewable energy generation is abundant and decrease consumption when there are generation constraints in the system (IRENA, 2019).

As electricity is a homogeneous good, which does not differ in quality, its price is the main determinant of the quantity demanded (Conover, 1984). As such, by modelling its price throughout the day, normal households may be presented with an incentive structure to change their consumption practices, similar to that of prosumers. The tariff schedules include the timing, duration and frequency of price signals (OFGEM, 2017). Time-based tariff structures can be static, determined in advance, or dynamic, determined based on the system conditions.

Figure 9. (IRENA, 2019). Types of ToU tariffs in the EU

Figure 9, lists and describes the different subtypes of ToU tariffs adopted by European countries.

Regardless of the tariff used, a number of benefits arise. Given the implicit demand response, consumers benefit through savings on their electricity bills. Additionally, the cost-reflective tariffs benefit suppliers while stimulating competition in the retail markets and acting as a driver for innovative business models (IRENA,2019)

(21)

Under static ToU tariffs, residential customers typically face two or more rates for their

electricity at fixed times of the day, referred to as a day/night differentiation. Often, the pricing follows the availability of renewable energy throughout the different time blocks. As such, the application of a static tariff would align the consumption practices of households with those of prosumers with PV installations.

Dynamic ToU (dToU) tariffs, on the other hand, are determined close to real-time based on actual system conditions. The potential for active peak reduction and load shifting brings benefits to stakeholders across the value chain (OFGEM, 2017). The flexible DR nature of household electricity consumption this tariff predisposes is closely aligned with that exhibited by smart prosumers. One important distinction arises, however, whereas smart prosumers rely on automation to manage their consumption given the market signals, households need to adjust their practices manually. As such, although the incentive structures may be closely aligned, households are expected to capture fewer benefits.

Studying the efficacy of the EC's PV mandate, given the policy objectives outlined in RPEU and assuming energy market participants provide the infrastructure and technology for smart prosumer integration, requires a comprehensive trial on aggregated smart prosumer behavioural patterns, given a set of exogenous shocks. Although a number of such trials have been conducted, data from them is not publicly available. However, the policy objectives can be translated into agent market behaviour through the use of dynamic price signals. As such, this behaviour can be approximated by using households' consumption data from dToU tariffs trials. Low price signals correspond to lower marginal cost of consumption, i.e. solar energy availability, incentivising supply following (SF) consumption in the given period. High price signals simulate consumption conditions under network constraint management (CM).

Price elasticity of demand can be estimated based on assumptions of rational decision making, utility maximization, and demand curve movement in a homogeneous goods oligopoly market structure (Oz Shy, 1996). Given these conditions, price movements are expected to have a causal inverse

relationship with electricity demand. As survey data has shown a general unwillingness of households to change non-cleaning appliances' time-of-use practices, the level of demand responsiveness to price is predicted to be consistent throughout the day (Klaassen, 2016). When price fluctuations occur often, agents are likely to exhibit response fatigue, requiring a relatively large change in utility before adjusting their consumption practices. As such, their demand functions for a given time block can be expected to exhibit a negative quadratic relation with price change (Mustapa et al., 2020). Lastly, assuming rational consumers seek to minimise the overall cost of consumption, the slope of the demand curve is

hypothesised to be steeper during Winter and Summer, as thermal regulation necessitates large amounts of electricity consumption, thus higher cost of use.

2.2 Low Carbon London:

"Low Carbon London – A Learning Journey" (LCL) was an integrated, large-scale, complex project measuring and evaluating the impact of a variety of low carbon technologies on London's electricity distribution network through a series of trials spanning a period of three years from January 2011 to December 2014 (Low Carbon London, 2016a).

(22)

London's electricity grid is unique due to the size and population density of the city. As a result, it has the highest concentrations of electricity demand, carbon-equivalent emissions production and the most demanding carbon reduction targets within the UK (Low Carbon London, 2016a). Further, the city also has the most promising scope for distributed generation, micro-generation, and electric vehicles (Low Carbon London, 2016a). The urban environment, combined with the degree of utilisation of its distribution network, predisposes high reinforcement costs to meet new demand. Through collaboration with the London School of Economics, LCL was designed in a way which would allow for the trial findings to carry relevance not just for London, but for the European continent in general (Low Carbon London, 2015).

The LCL Smart Meter trial was facilitated by utility provider EDF Energy. The company installed 5,533 meters within the houses of recruited customers. To ensure that an even grouping of socio- economic prospects was recruited onto the trial, representative of the London-residential customer mix, an analysis was undertaken to map all definitive prospects to the corresponding ACORN groups (Low Carbon London, 2016b). ACORN data is a customer classification system which segments the UK

population using demographic data and social factors. In accordance with this demographic distribution, customers were invited to participate in a dynamic Time of Use tariff (dToU) trial. 1,119 customers were recruited onto the dToU tariff. The final composition of the non-ToU and dToU trial population

demographic can be seen in Figure 10.

Figure 10. (Low Carbon London, 2016b). Population groups ACRON distribution

Most ToU tariff initiatives in GB have, to date, been led by energy suppliers and based on potential wholesale cost savings rather than on the potential network benefits (Hledik, 2017) (Low Carbon London, 2016a). In contrast, due to the network characteristics and the expected needs of London, the LCL dToU trial was targeted at alleviating network stress. This was done through the use of constraint management and supply following tariffs. A high ToU price represents the need for CM and encourages customers to shift load away from periods where networks are stressed. Displaying a Low-

(23)

High-Low (LHL) price pattern, customers could shift demand to a low rate period on either side of a high rate period. SF events included periods of low prices of 3, 6, or 12 hours in length. These durations were set in accordance with the results of system-wide wind generation data analysis. A combination of both CM and SF events was used to encourage customers to relocate consumption from periods with shortfalls in the supply of power to periods where there is surplus, e.g. periods of high renewable generation (Low Carbon London, 2016a).

Customers were informed a day in advance before the scheduling of a dynamic price change (Low Carbon London, 2016b). Messages were communicated to all households via In-Home Displays (IHDs) and, if opted into, by text message. Customers were issued high (67.20p/kWh), low (3.99p/kWh), or normal (11.76p/kWh) price signals and the times of day these applied. All non-Time of Use customers were on a flat rate tariff of 14.23p/kWh.

2.3 Relevant Data:

For the analysis, the measured data spanning one year, from January 1st, 2013, to December 31st, 2013, of both the control and treatment groups is used. The study's timeframe is considered large enough to include seasonal fluctuations and assess structural changes in behaviour and response fatigue (Klassen et al., 2016). Although initially 1119 dToU and 4414 non-Tou consumers were recruited to form the trial composition, due to a consistent lack of observations of some participants, the sample size was limited to 𝑁𝑁𝑇𝑇𝑇𝑇𝑇𝑇 = 1111 and 𝑁𝑁𝑛𝑛𝑇𝑇𝑛𝑛−𝑇𝑇𝑇𝑇𝑇𝑇 = 4397.

The smart meters collect electricity consumption data by measuring and recording household consumption in kWh every half an hour. A separate dataset contains the dates, times, and types of the tariffs scheduled. As a consequence of showing the dynamic tariff on the IHDs and informing

participants via SMS, given the rational decision-making assumption, consumption is expected to be higher during low-priced periods and lower during high-priced periods. To assess the presence of flexible DR, the mean load of the treatment group is compared to that of the reference group. As the ACORN composition of both groups is similar, the households' specific circumstances affecting electricity consumption are also assumed to be similar.

2.4 Methodology:

Several different regression designs were considered. Firstly, Difference-in-Difference's ability to create a counterfactual from treatment and control groups data, subsequently used to estimate a causal effect, was appraised. However, as the parallel trends assumption cannot be confirmed and participants in the treatment group were self-selected, this approach would lead to biased results. Further, an IV OLS fixed effects regression was also considered. Although given an appropriate instrument and control variables, this approach would lead to the most robust results, it would be computationally arduous for Stata to run such a regression given the size of the data file (For more information, see Appendix).

The methodology used in this paper is replicated from Klassen et al. (2016) study of residential responsiveness to dynamic tariffs, based on a large field test in the Netherlands. As the electricity consumption of each household is assumed to be an independent identically distributed random

(24)

variable and the sample size is sufficiently large, the central limit theorem implies that the mean of household load is a normally distributed random variable centred on the true mean (Klassen et al., 2016). As such, the standard deviation can be calculated using the variance. To quantify the magnitude of the flexible DR, the average load of the treatment group is compared to that of the reference in two 8-hour time periods: Period 1 (09:00-17:00), a time block with relatively high system-wide renewable energy generation, and Period 2 (17:00-01:00), a time block with relatively high system constraints.

Further, to assess where the DR responsiveness is dependent on the prevalence of cooling and heating degree days, loads of the two groups are compared throughout the different UK meteorological seasons as defined by the UK Met Office (Met, ND).

To assess a significant difference between the load of the participants and the reference group a two-sample t-test is conducted, using the following input:

𝑡𝑡 = 𝐶𝐶��������−𝐶𝐶𝑇𝑇𝑇𝑇𝑇𝑇 ��������������𝑛𝑛𝑇𝑇𝑛𝑛−𝑇𝑇𝑇𝑇𝑇𝑇

(𝑁𝑁𝑇𝑇𝑇𝑇𝑇𝑇−1)𝑠𝑠𝑇𝑇𝑇𝑇𝑇𝑇2 +(𝑁𝑁𝑛𝑛𝑇𝑇𝑛𝑛−𝑇𝑇𝑇𝑇𝑇𝑇−1)𝑠𝑠𝑛𝑛𝑇𝑇𝑛𝑛−𝑇𝑇𝑇𝑇𝑇𝑇2

(𝑁𝑁𝑇𝑇𝑇𝑇𝑇𝑇+𝑁𝑁𝑛𝑛𝑇𝑇𝑛𝑛−𝑇𝑇𝑇𝑇𝑇𝑇−2) ∗�𝑁𝑁𝑇𝑇𝑇𝑇𝑇𝑇1 +𝑁𝑁𝑛𝑛𝑇𝑇𝑛𝑛−𝑇𝑇𝑇𝑇𝑇𝑇1

(1)

s.t.

𝐶𝐶𝑇𝑇𝑇𝑇𝑇𝑇𝚤𝚤

�������� = (𝑡𝑡2−𝑡𝑡1 1)∑ 𝐶𝐶𝑡𝑡𝑡𝑡21 𝑇𝑇𝑇𝑇𝑇𝑇𝑖𝑖𝑡𝑡 ∀𝑖𝑖 ∈ {1 … 𝑁𝑁𝑇𝑇𝑇𝑇𝑇𝑇} (2) 𝐶𝐶𝑇𝑇𝑇𝑇𝑇𝑇

������� = 𝑁𝑁1

𝑇𝑇𝑇𝑇𝑇𝑇𝑁𝑁𝑖𝑖=1𝑇𝑇𝑇𝑇𝑇𝑇𝐶𝐶�������𝑇𝑇𝑇𝑇𝑇𝑇𝚤𝚤 (3)

𝐶𝐶𝑛𝑛𝑇𝑇𝑛𝑛−𝑇𝑇𝑇𝑇𝑇𝑇𝚤𝚤

������������� = (𝑡𝑡2−𝑡𝑡1 1)∑ 𝐶𝐶𝑡𝑡2 𝑛𝑛𝑇𝑇𝑛𝑛−𝑇𝑇𝑇𝑇𝑇𝑇𝑖𝑖𝑡𝑡

𝑡𝑡1 ∀𝑖𝑖 ∈ {1 … 𝑁𝑁𝑛𝑛𝑇𝑇𝑛𝑛−𝑇𝑇𝑇𝑇𝑇𝑇} (4)

𝐶𝐶𝑛𝑛𝑇𝑇𝑛𝑛−𝑇𝑇𝑇𝑇𝑇𝑇

������������� = 𝑁𝑁 1

𝑛𝑛𝑇𝑇𝑛𝑛−𝑇𝑇𝑇𝑇𝑇𝑇𝑁𝑁𝑛𝑛𝑇𝑇𝑛𝑛−𝑇𝑇𝑇𝑇𝑇𝑇𝐶𝐶�������������𝑛𝑛𝑇𝑇𝑛𝑛−𝑇𝑇𝑇𝑇𝑇𝑇𝚤𝚤

𝑖𝑖=1 (5)

𝑠𝑠𝑇𝑇𝑇𝑇𝑇𝑇2 =𝑁𝑁 1

𝑇𝑇𝑇𝑇𝑇𝑇−1𝑁𝑁𝑖𝑖=1𝑇𝑇𝑇𝑇𝑇𝑇(𝐶𝐶������� − 𝐶𝐶𝑇𝑇𝑇𝑇𝑇𝑇𝚤𝚤 ������)𝑇𝑇𝑇𝑇𝑇𝑇 2 (6)

𝑠𝑠𝑛𝑛𝑇𝑇𝑛𝑛−𝑇𝑇𝑇𝑇𝑇𝑇2 =𝑁𝑁 1

𝑛𝑛𝑇𝑇𝑛𝑛−𝑇𝑇𝑇𝑇𝑇𝑇−1𝑁𝑁𝑛𝑛𝑇𝑇𝑛𝑛−𝑇𝑇𝑇𝑇𝑇𝑇(𝐶𝐶������������� − 𝐶𝐶𝑛𝑛𝑇𝑇𝑛𝑛−𝑇𝑇𝑇𝑇𝑇𝑇𝚤𝚤 ������������)𝑛𝑛𝑇𝑇𝑛𝑛−𝑇𝑇𝑇𝑇𝑇𝑇 2

𝑖𝑖=1 (7)

(25)

Where 𝐶𝐶�������� and 𝐶𝐶𝑇𝑇𝑇𝑇𝑇𝑇𝚤𝚤 ������������� are the average loads (kW) of each individual participant and 𝑛𝑛𝑇𝑇𝑛𝑛−𝑇𝑇𝑇𝑇𝑇𝑇𝚤𝚤

reference household 𝑖𝑖, during a certain individual time period, expressed by the subscripts t1 and t2. 𝐶𝐶𝑇𝑇𝑇𝑇𝑇𝑇

������� and 𝐶𝐶������������� and (C_(non-ToU) ) ̅ are the average loads of both the participant and reference 𝑛𝑛𝑇𝑇𝑛𝑛−𝑇𝑇𝑇𝑇𝑇𝑇

group, of which the sample size is expressed by NToU and Nnon-ToU, respectively. The unbiased sample variance of both groups is indicated by 𝑠𝑠𝑇𝑇𝑇𝑇𝑇𝑇2 and 𝑠𝑠𝑛𝑛𝑇𝑇𝑛𝑛−𝑇𝑇𝑇𝑇𝑇𝑇2 . The t-value is compared against the critical value defined by the t-distribution to determine the significance. If the p-value is less than 5 per cent, the difference between 𝐶𝐶������� and 𝐶𝐶𝑇𝑇𝑇𝑇𝑇𝑇 ������������� is considered significant. 𝑛𝑛𝑇𝑇𝑛𝑛−𝑇𝑇𝑇𝑇𝑇𝑇

2.5 Results and discussion:

The results section is subdivided into three subsections, covering the results of the DR program during CM and SF tariffs and a discussion on their implications within the wider household consumption framework.

2.5.1 Constraint Management tariff:

Under the assumption of a negative quadratic relation with the price change, and given that the

"High" tariff is nearly six times larger than the "Normal" tariff, a relatively large consumption contraction was theorized. Indeed per the null hypothesis, the two-sample t-test demonstrates a significant

decrease in the load of the treatment group in both periods. For Period 1, a 15.2% decrease in load size was measured when comparing the treatment versus control groups, with 𝐶𝐶�������=18.27 W, 𝑇𝑇𝑇𝑇𝑇𝑇

𝐶𝐶𝑛𝑛𝑇𝑇𝑛𝑛−𝑇𝑇𝑇𝑇𝑇𝑇

�������������=21.53 W, t(343099)= -49.52, p-val<0.001. For Period 2, am 11.3% decrease in consumption was calculated, with 𝐶𝐶�������=27.92 W, 𝐶𝐶𝑇𝑇𝑇𝑇𝑇𝑇 �������������=31.45 W, t(803168)= -64.03, p-val<0.001. The results 𝑛𝑛𝑇𝑇𝑛𝑛−𝑇𝑇𝑇𝑇𝑇𝑇

show a significant willingness to decrease net consumption during periods of constraint in the wider system. In contradiction to the inference of constant price elasticity of demand throughout the day, however, the data indicates that households' disposition for consumption reduction is nearly a third greater during the morning and mid-day hours rather than in the afternoon and at night.

In the Winter and Summer seasons, for both Periods 1 & 2, in response to high tariffs, dToU households shrank consumption by 11.7%, in comparison to their non-ToU counterparts, with 𝐶𝐶������� 𝑇𝑇𝑇𝑇𝑇𝑇

=25.68 W, 𝐶𝐶�������������=29.08 W, t(617890)= -55.46, p-val<0.001.In the Spring and Autumn months, for 𝑛𝑛𝑇𝑇𝑛𝑛−𝑇𝑇𝑇𝑇𝑇𝑇

the same time periods, consumers part of the treatment group consumed 12.6% less electricity, with 𝐶𝐶𝑇𝑇𝑇𝑇𝑇𝑇

�������=24.39 W, 𝐶𝐶�������������=27.89 W, t(523356)= --56.01, p-val<0.001. Contrary to the null hypothesis, 𝑛𝑛𝑇𝑇𝑛𝑛−𝑇𝑇𝑇𝑇𝑇𝑇

the higher prevalence of heating degree days (HDD) and cooling degree days (CDD) within the Summer and Winter seasons did not affect the willingness of households to shift thermal regulation loads.

2.5.2 Supply Following tariff:

During the LCL trial, the "Low" tariff was priced around two and a half times lower than the

"Normal" tariff. Given the assumed functional form specification of household consumption, load changes were expected to be statistically significant yet smaller in magnitude, as the level of price change is less than half of that of the "High" tariff. In Period 1, households under the dToU tariff

(26)

consumed 1.6% more electricity than their static tariff counterparts, with 𝐶𝐶�������=24.29 W, 𝑇𝑇𝑇𝑇𝑇𝑇

𝐶𝐶𝑛𝑛𝑇𝑇𝑛𝑛−𝑇𝑇𝑇𝑇𝑇𝑇

�������������=23.91 W, t(1.0e+06)= 8.226, p-val<0.001. In Period 2, electricity load was 3.1% higher, with 𝐶𝐶𝑇𝑇𝑇𝑇𝑇𝑇

�������=27.37 W, 𝐶𝐶�������������=26.52 W, t(653423)= 14.32, p-val<0.001. The results suggest that, although 𝑛𝑛𝑇𝑇𝑛𝑛−𝑇𝑇𝑇𝑇𝑇𝑇

households are willing to shift loads in abundant renewable energy generation periods, the magnitude of the shifted consumption is relatively small. Excess mean consumption was almost twice as high in Period 2 compared to Period 1. This observation, however, may be influenced by the general pattern of residential electricity use, as it is skewed towards the afternoon and night periods. Nonetheless, based on the econometric findings, it can be concluded that even under CM and SF tariffs, households' preference to consume in the afternoon cannot be altered.

In Summer and Winter, for both Period 1 & 2, under a “Low” tariff, consumption increased with only 0.3%, with 𝐶𝐶�������=26.46 W, 𝐶𝐶𝑇𝑇𝑇𝑇𝑇𝑇 �������������==26.36 W, t(92796)=1.51, p-val=0.0653. However, as the p-𝑛𝑛𝑇𝑇𝑛𝑛−𝑇𝑇𝑇𝑇𝑇𝑇

value is higher than 5 per cent, the results are considered statistically insignificant. In Spring and Autumn, for the same periods, mean load was 1.2% higher, with 𝐶𝐶������� =23.71 W, 𝐶𝐶𝑇𝑇𝑇𝑇𝑇𝑇 �������������==23.43 W, 𝑛𝑛𝑇𝑇𝑛𝑛−𝑇𝑇𝑇𝑇𝑇𝑇

t(719115)=5.4576, p-val<0.001. Given these results, it cannot be concluded that households optimize their respective electricity bills by shifting electricity-intensive thermal regulation to periods of relatively low prices.

2.5.3 Discussion:

Dynamic ToU tariffs simulate the financial incentives structure smart prosumers would be exposed to, however, they do not encapsulate the full range of intrinsic and extrinsic factors affecting household consumption practices. Further, although the results discussed above may be internally valid, the broader conditions of the electricity market have changed dramatically since the trial and will continue to change until the regulation is implemented in 2029. As such, to provide external validity of the findings and to allow for their extrapolation towards the estimation of the efficacy of the mandate in relation to the policy objectives, the results have to be situated contextually.

Based on the residential electricity consumption framework developed in Section 1, households' finances can be expected to be more sensitive towards their respective electricity consumption

practices, as both demand and price are projected to grow. As previously outlined, the average share of electricity in final energy consumption is expected to increase by close to a third by the time of the mandate implementation, from 23 per cent today to 30 per cent by 2030 (JRC, 2020). Additionally, the value of European Power Benchmark, as of Q4 2021, has increased 400 per cent year-over-year and 85 per cent quarter-over-quarter (EC, 2022a). Although this growth is in large part due to the pandemic and ongoing war, prices of electricity are also highly correlated with inflation and the supply of energy goods, which are globally determined and, as such difficult for the EU to manage individually. Further, inflation also reduces the purchasing power of agents, requiring them to allocate a higher proportion of their income to necessity goods (McKinsey, 2021). The combination of these factors predisposes a higher relevance imposed by rational households on their energy practices. As such, they're further incentivized to shift their respective loads based on the cost of consumption in a given time block. This suggests that the quantitative findings underestimate the expected magnitude of flexible demand response.

Referenties

GERELATEERDE DOCUMENTEN

License: Licence agreement concerning inclusion of doctoral thesis in the Institutional Repository of the University of Leiden Downloaded.

Voor zon moet je gewoon veel meer opslag hebben maar goed daar wordt aan gewerkt.. I: We hebben het al heel lang over de

The multi-level perspective gives insight in what kind of actors are interesting for this the- sis, namely regime level actors: involved in tactical governance

Ministry of Foreign Affairs of the Netherlands, Focus Letter Development Cooperation (The Hague: Dutch Government, 2011): 3.; Ministry of Foreign Affairs of the Netherlands, A World

[r]

In order to investigate this relationship, I reduce my panel data to the country level and use the capital share of zombie firms per country per year as the dependent variable..

Keywords: SMEs, Open innovation strategy, knowledge exploration (inbound open innovation), knowledge exploitation (outbound open innovation), resource-based view, open

We focus on the wealth (financial distress, making ends meet, financial situation and economic situation), saving behavior and beliefs (income uncertainty,