Analysing the potential of citizen-financed community renewable energy to drive Europe's low-carbon energy transition

In 2018, the real amount invested in the European Union’s energy transition fell short of the funding level required to reach the 2030 climate and energy targets by €179 billion. Citizen-led finance in renewable energy development emerges as an innovative tool to bridge this investment gap. However, in spite of the European Union’s ambition to involve local communities for co-driving the low-carbon energy transition, there is no comprehensive analysis quantifying citizens’ potential to co-finance and participate in community renewable energy initiatives. We address this knowledge gap through a representative choice experiment survey that collected responses to 389,640 hypothetical investment options on renewable energy schemes across all European Union Member States, and estimate the social potential of European citizens to participate and invest in community-administered wind farms. Results from a novel survey-based social simulation indicate that €176 billion could be obtained from citizen-led finance in community-administered wind farm developments, enough to halve the investment gap to achieve a 32% RES share by 2030. Reaching this potential would lead to the deployment of 91GW of installed wind power capacity, generate up to 196 TWh of renewable energy annually across Europe, and trigger an average increase of 8.3% in final renewable energy consumption. This would lead to the abatement of over 103 MtCO 2 eq annually and result in a 2.4% annual reduction in greenhouse gas emissions from 2018 levels. Our analysis substantiates the case for easily accessible, risk-insured community investment options across Europe to unlock citizens’ social potential for investing in community renewable energy.


1
Analysing the potential of citizen-financed community renewable energy to drive Europe's low-carbon energy transition. 1

. Introduction
Launched in 2015, the Energy Union is the EU's largest and most ambitious climate and energy legislative effort to fully decarbonise Europe's energy system by 2050, with a set of intermediate energy and climate targets for 2020 and 2030. These now reflect a 32% share of RE in final energy consumption, a 32.5% improvement in energy efficiency, and a 40% reduction in greenhouse gas (GHG) emissions from 1990 baseline levels [1]. However, tangible policy instruments and implementation mechanisms foreseen under current National Energy and Climate Plans (NECPs) for the period 2021-2030 remain elusive and inadequate for reaching the abovementioned targets [2], [3]. In light of insufficient efforts, a viable pathway for decarbonisation at the extent necessary to reach climate neutrality by mid-century remains elusive.
EU Member States (MSs) must therefore explore viable GHG emission reduction measures to successfully decarbonise their economies and realise an increasingly steep emissions reduction pathway to operate within an EU carbon budget of around 96 GtCO2 for the period 2010-2050 [4].
The investment requirements to realise such emission reductions are substantial and currently not being met [5]. Estimates indicate that about €380 billion are required annually over the next 10 years in order to achieve the EU's 2030 climate and energy targets, nearly double the 2018 investment level of €201 billion [5]- [7]. This resulted in an investment gap of €179 billion for 2018. Estimates further project that no less than 9% of the foreseen annual investments, at least €34 billion per year, will have to finance the deployment of renewable generation capacity to reach a 32% share of the EU's final energy consumption by 2030 [5]. This translates into a cumulative investment of €340 billion in RE capacity over the 2019-2030 period.
The EU aims to bridge the resulting investment gap through market-driven strategies that place citizens at the core of the Energy Union by having them "[…]take ownership of the energy transition, benefit from new technologies to reduce their bills, [and] participate actively in the market" [8, p. 2]. In that respect, citizen participation in community-based RE (CRE) generation through collective investment and shared ownership schemes, emerges as an innovative tool to unlock citizens' social potential to contribute in bridging the existing investment gap, as well as their GHG emission abatement potential through the co-generation and local sourcing of clean energy. This positions CRE as an important vehicle through which bottom-up, community-based climate mitigation actions can occur, and can thereby empower individual citizens to contribute towards Europe's objective of carbon neutrality by mid-century. Furthermore, CRE schemes can generate ancillary benefits for its owners and surrounding communities and thereby contribute in increasing social acceptance of clean energy alternatives [9]- [18]. Ancillary benefits may include additional sources of income from electricity sales or through dividends from the ownership of shares and/or land rent [12], [19]- [21], lower energy costs derived from local or self-consumption [22], enhanced social cohesion and sense of community [23]- [25], and increased environmental awareness and stewardship [26], among others.
Importantly, CRE schemes can have an additional positive characteristic above selfownership schemes. Firstly, by distributing the initial capital investment needed throughout a large group of small-scale investors, collective investment schemes in CRE lower the required investment amount for individuals vis-a-vis a fully self-owned system [13], [17], [18], [27], [28].
This can allow for under-privileged groups, such as lower income households or those who lack the proper infrastructure for a self-owned RE installation, to participate in RE generation. Along with this added benefit, citizen participation in CRE decreases social indifference and uncertainty towards new RE installations, an effect that can increase the social acceptance for new RE developments [29]- [33] and therefore expedite the low-carbon energy transition.
However, despite the documented socio-economic benefits attributed to CRE, no efforts have been conducted to quantifyyet alone monetise -individual citizens' potential contribution to financially participate in CRE initiatives. Furthermore, no research has yet aimed to translate different levels of financial participation in CRE into GHG emission reductions. The only attempt to approach such a quantification comes from [34], who use country-level data for estimating the potential number of RE prosumers (consumers and producers of RE) in Europe across a broad base of arrangements including self-ownership, CRE, public ownership and firm adoption. Assuming a 100% RE scenario, the authors estimate that about 113 million households could become prosumers across the EU by 2050. However, the study does not fully consider the participation decisions of households to either adopt or not a RE technology, and further assumes that a hypothetical financial investment per household is a function of average savings rates alone 1 . 1 [34]state their estimation method as follows: "The Eurobarometer contains statistics on which topics concern citizens the most. By selecting the relevant topics, the share of households that will want to invest in renewable energy is estimated. The total amount this group can invest is limited by multiplying their normalized average savings rate between 1995 and 2015 with an estimated minimal and maximum amount each household will invest yearly" (p. 13). No further information on this methodology is given in their report.

5
This study aims build on the work of [34] by investigating the optimal combination of different financial attributes and operational configurations that maximise individual citizen investments in community-based RE generation schemes. We quantify the resulting investment volumes and aggregate them across the EU-28 to estimate European citizens' aggregated financial participation in CRE, defined as the social potential. We then assess the extent by which the EU's social potential can effectively finance the EU's 2030 RE target, and estimate the volume of RE generated and GHG emissions reductions stemming from it.
In order to do so, we use data from a choice experiment conducted across all EU-28 where respondents are presented with different investment options to co-finance RE schemes with different characteristics and varying attributes. Using imputed choice probabilities derived from multinomial logit estimates on the choice experiment data, we compute the social potential as the expected investment from the representative citizen of each EU MSs; we call this process Survey-Based Social Simulation (SBSS). Section 2 introduces the choice experiment as our main data collection tool. Section 3 describes the development of the SBSS methodology. In section 4 we present the results stemming from our analytical procedure and assess the potential of our estimated citizen financial participation in CRE to effectively bridge the current investment gap for financing the RE capacity needed to reach the EU's 2030 RE target. We then contextualise our results within the EU's broader climate action commitments by translating our estimated social potential into GHG emission reductions; and close by highlighting limitations and caveats to our methodology. In section 5, we explore the effects of existing RE subsidies on our estimated social potential, and reflect on the policy implications of our results. Section 6 concludes.

Data collection
The main data utilised for this study was obtained from the responses to a Choice Experiment (CE) conducted as part of an international online survey with private citizens across 31 countries (EU-28, Norway, Switzerland, and Turkey). The internet-based survey was distributed to about 600 respondents in each country through a random sampling procedure that relied on email panels, with a total final sample of 16,235 respondents. Respondents were recruited using a compensation mechanism of at least €5 per person to incentive participation and ensure that not only people with energy interests made up the sample. As shown in table A.1 (see Appendix), quotas were set from national sociodemographic indicators pertaining to age, gender, 6 and income levels in order to ensure the representability of the samples for all 31 countries. To frame our study in the context of EU climate and energy targets, we drop sample responses from Norway, Switzerland and Turkey.
The purpose of the CE was to identify respondents' levels of interest in participating in a community-based investment scheme to finance solar or wind projects, and to investigate what specific set of investment attributes of RE initiatives, including financial and operational conditions, drive citizen participation. Survey respondents were presented with eight different choice scenarios, each one displaying a total of three options to choose from: two hypothetical investment opportunities (option A, option B), and a third "opt-out" option (option C) provided in the case where a respondent had no interest to invest. Respondents were asked to choose which of these three options they would prefer if confronted with the same situation in real life. The choice scenarios were built following a D-efficiency criteria [35], and organized into three blocks of eight scenarios. Each respondent was presented the scenarios from a randomly selected block in a random order 2 , and asked to pick one option for each of the eight different choice scenarios. This As illustrated in figure 1 and specified in table 1 below, each choice scenario included two scenario-specific characteristics that applied to all three choice options (A, B, and C) within a given scenario. In addition, the choice scenarios included four option-specific attributes that varied between choice options A and B, with option C as the default opt-out option where all attribute variables are set to zero. Figure A.1 in the Appendix offers an example of one of the eight choice scenarios utilised in the CE and as shown to English-speaking survey respondents, along with a descriptive script introducing the CE exercise and explaining the scenario and premise of the investment options. 2 For a full description of the choice experiment design, testing and results please see [27].     4 . This initial assessment points toward a potential trade-off between installed capacity and investor acceptance levels (as manifested by a lower willingness to co-invest in CRE initiatives), indicating an inversely proportional relationship between both variables.
It further suggests that collective investments on RE initiatives are seen favourably as viable financial instruments in nations where public acceptance issues accompanying new energy infrastructure are not yet strongly rooted [38]. Alternatively, countries manifesting low interest may be illustrating small-scale energy investment constraints due to prohibitive capital investment requirements [39], [40]. This may be further exacerbated by a combination of stricter spatial planning criteria [39], [41], lengthy permitting procedures and legal disputes [42], [43], and increasingly stringent RE compensation mechanisms [44], particularly in countries with longstanding traditions on cooperative association such as Denmark and Germany.

Methodology
CEs are a well-suited data generation method for modelling and interpreting respondents' choice dynamics using probability-based discrete choice models to account for people's preferences and decision-making processes [45]- [49]. CEs are grounded in economic random utility theory, which assumes that the utility level person n experiences from choice option i can be separated into one observable V(.) and a random ἐ(.) component [46]. In our analysis, both components are a function of the option-specific attributes Zin, and the scenario-specific characteristics Sn (both illustrated in figure 1 and described in table 1). As such, the latent utility Uin from any choice option can be expressed as: Based on equation (1), we can assume that any choice option i will be preferred/selected over some Mathematically, we define this individual (per person) social potential as in (2). In what follows we substitute the subscript n for N, as we calibrate the SBSS method on the representative individual of each nation N, as opposed to individual respondents (n).
In order to calculate the individual social potential ( ) we first impute the probability (PN) that the representative individual in country N chooses an investment option and then multiply this value by the capital requirement (bN) that the individual is asked to contribute to for cofinancing an RE installation. This yields the expected investment per citizen of country N as expressed in (2).
Equation (3) defines the probability PN that the representative individual of country N will invest in an option with a capital requirement bN as a result of he/she ascribing a higher utility U to any one investment option (A or B) than to opting-out (option C).
The computation of choice probabilities, PN, is accomplished by imputing the average probability of choosing option A or B from discrete choice multinomial logit (MNL) models estimated on the CE data. The imputations are based on a choice option with the feasible, preferred configuration of attributes defined in Step 1 below. Country-specific social potentials provide the main building block to then proceed to the second stage of our analytical exercise, where we quantify the RE generation and GHG abatement potential stemming from citizen-financed CRE schemes across the EU.

Estimating the social potential
Employing the SBSS entails the use of choice model parameter estimates to impute choice probabilities given the effects of different combinations between option-specific attributes and scenario-specific characteristics, and maximizing some objective function subject to these imputed probabilities. This allows us to estimate both the probability that any one specific option under any given scenario is selected instead of competing options, and the changes in this probability given some specified modifications in the design settings. Assuming the latent utility in (1) is linear with respect to the explanatory variables and that the error term takes a Gumbel extreme value distribution, leads us to the adoption of the alternative specific multinomial logit model (MNL) in (4)the common workhorse model for discrete choice frameworks.
As illustrated in the conceptual design of our CE (figure 1), the MNL models the probability of a discrete choice outcome as a function of the option-specific set of attributes and the scenario-specific characteristics . The effects of the variables within Zin and Sn on the latent utility levels Uin of individual respondent n are represented by the vectors β and αi. Since we are interested in the decision to invest vs. not invest, we set option C (opting out) as the base alternative, and assign a value of zero to its corresponding coefficient vector . This allows to more easily interpret the output of the MNL as the effects (β and αi) that the different variables from option and scenario-specific attributes ( , ) have on the probability of opting-out from an investment option (A or B) presented in the choice scenario.

Step 1: Estimating the effects of choice attribute variables on choice probabilities
Equation (4)

Step 2: Imputing the probability of accepting an investment option
Following from step 1, we input the most preferred attribute levels identified and defined in Step 1 into equation (4), and compute the probability PN of accepting the most preferred investment configuration outlined above. Important to note that up until this stage we have not yet considered the other critical quantity for estimating the individual social potential ( ); that is, the average capital invested per person (bN). We show this formally in equation (5)

Step 3: Maximising the expected funds collected per country
The outcome of step 2 above allows to now calculate the optimal investment requirement to ask from the representative individual in each nation N based on the most preferred variables inputted in the simulation of the MNL model. We thus define ̃ as the country N's 'social potential' for investing and participating in community-administered wind farm cooperatives, with indicating country N's population with a reasonable expectation to invest (ages 25-64) 7 .
In spite of our goal to maximise ̃, for now we ignore the population component and return to from (2), which considers only the representative individual from each sampled nation. We thus face a two-variable maximization problem, where we aim to maximise (2) with respect to the representative probability PN and capital requirement bN, yet constrained by the relationship in (5).
Substituting (5) into (2) we arrive at (7): Taking the first derivative of (7) with respect to bn we obtain (8): By setting (8) equal to zero and solving for bN we obtain an analytical solution for the optimal capital requirement (expressed as * ) that maximises the expected capital offering ̃ that the representative individual from country N is willing to provide 8 : * = −2 (̂1 +̂2 ) The optimal capital requirement * is input back into equation (5) and solved in order to obtain the probability * that a representative citizen of country N chooses to invests capital level * into a community-administered wind farm, with a 20-year holding period and a market-based profit rate. Inputting * and * into (6) and solving results in the final estimation of the social potential (̃) for investing and participating in community-administered wind farms in each nation N.

Quantifying the RE generation and GHG abatement potentials
The second stage in our analytical procedure utilises the calculated social potential of each MS as the starting point to quantify the GHG abatement potential of individual citizens across the EU. In order to do so, we first quantify the installed wind power capacity (GW) that could be bought with the social potential of each MS derived from the total volume of individually committed investments in community-administered wind farms. This is done by dividing the expected volume of funds collected per country (the social potential) by the European averaged total installed generation capacity costs for wind power technology (for this, we use €1,939.13/kW as the value derived from market data described in Step 1). With this, we obtain the installed wind power capacity (in MW) that could be purchased with the funds collected from the social potentials of each MS under current market conditions.
Country-specific wind power capacities are then combined with national energy productivity ratings to quantify the RE generated annually from the installed wind power capacity obtained for each MS and across the EU. Building on this calculation, we then input the RE generated into each country's 2017 gross final energy consumption serviced by RES in order to quantify the (percentage) increase in the share of RE consumption within each country's total gross final energy consumption (see Table A

The most preferred investment attributes
Following from Step 1, we report the results of an initial estimation of the MNL in (4) Table 3 shows that, on average, European citizens strongly prefer a community-owned legal entity (e.g. energy cooperative) for administering the RE installation they invest in over government or utility company administrators, and slightly prioritise a company-managed RE installation before a government-administered alternative. In spite of high heterogeneity, we observe a slight overall inclination for wind farms over solar parks as the preferred technology to invest in; this is likely highly region-specific as explored in [27]. European citizens are also more willing to invest if they see the RE installation from their household. As expected, higher profit rates make the investments more preferable. Specifically, for every additional €100 obtained as profit, we observe a

The social potential for wind farm cooperatives in the EU-28
Estimating CE participants' responses to the optimal investment scenario identified above, and combining these with country-specific profit rates as shown in table A  The imputed probability that a given respondent accepts the ideal investment option is of 20% across the entire sample. With the caveat that this average probability is not weighted by country population sizes, this result suggests that about one in five European citizens would be willing to invest in a feasibly configured community-administered wind farm development. We note the heterogeneity in the expected collection per citizen (π N ), where generally nations with lower wind power capacities show higher acceptance rates. This corresponds with the descriptive results obtained from the CE responses (table 2), which we interpret as an increased interest from citizens in EU MSs with low wind power capacities to have access to low-risk, trustworthy investment options in this technology.

Bridging the investment gap
As shown in figure 2 below, current estimates indicate an investment gap of €179 billion annually to achieve the EU's 2030 climate and energy targets [1], [6]. No less than 9% of the foreseen annual investments over the next 10 years (2020-2030) -at least €34 billion annuallywill have to finance the deployment of RE generation capacity in order to reach a 32% share of the EU's gross final energy consumption by 2030 [5]. This translates into a cumulative investment of €340 billion over the next decade, and positions the EU's social potential of €176 billion for As depicted in figure 2 above, the social potential of €176 billion that European citizens could contribute with through collective investment schemes in CRE respond directly to this need.
When evenly distributed throughout the 10-year timespan mentioned above, they result in an annual investment of €17.6 billion, enough to halve the investment requirements foreseen to achieve a 32% RES share by 2030. In light of this huge potential, the EU's energy-related carbon mitigation efforts could greatly benefit from the proactive financial participation and involvement of European citizens. Policies that reach out to and unlock this potential are therefore desirable and should be carefully considered for a timely, cost-effective, and participatory implementation of a low-carbon energy system. This is further explored in section 5 below.

GHG abatement potential from citizen-financed wind farms
Following the process outlined in section 3.2, the EU's social potential for co-financing wind farms would be sufficient to "purchase" a total of 90,900 MW of wind power capacity across the entire EU. This represents a larger volume than the national electricity generation capacities of 24 different MSs [51].
When multiplied by their corresponding national wind energy productivity ratings, the installed wind power capacities from each country's social potential yield the final RE (in GWh) generated annually by each MS and across the EU. Building on this initial calculation, the RE EU social potential €17.6 billion (52%) Investment gap €179 billion (47%)

Investments for 2030 RES target €34 billion (9%)
Current total investments €201 billion (53%) generated is input into each country's 2017 gross final energy consumption serviced by RES in order to quantify the increase that the share of RES consumption would experience in each country's total gross final energy consumption if the social potential were realized. This is illustrated in figure 3 below.  [53]. See Table  A.5 in the Appendix for details).
Results indicate an average 8.3% increase in the consumption of RES across the EU when the social potential for co-financing community-administered wind farms is included. As shown in Table A.5 (Appendix), this translates into a total of 196 TWh of additional energy consumed.
Assuming such consumption does not add tobut instead substitutes -196 TWh of energy consumed from conventional energy sources, the GHG emissions that could be potentially abated amount to 103.4 MtCO2-eq annually (table 5 below). This represents a 2.4% reduction in annual emissions from 2018 EU aggregate levels 12 and over 3% of the GHG emissions stemming from the energy sector in 2018 [54]. While this reduction in emissions is a substantial improvement expediting the projected pace of emission reductions, it is by no means sufficient to put the EU on track to achieve its 2030 GHG reduction target. The EU would still need to abate an additional 274 MtCO2-eq per annum to achieve a 40% reduction by 2030 [55]. The EU should therefore adopt 11 UK's 2030 RES target not reported in its National Energy and Climate Plan (NECP) nor submitted to the EU. 12 Excluding international aviation/shipping and LULUCF emissions, including indirect CO2 emissions. additional carbon reduction measures in order to successfully decarbonize its economy and realise an emissions reduction pathway sharp enough to operate within an 'EU carbon budget' of around 96 GtCO2 per annum [4]. Table 5. Analysis of the annual GHG abatement potential from realizing the social potential for wind-farm cooperatives (own calculations with input data from [56] as described in Section 3.2).

Country
Energy generated annually (GWh)

Limitations and caveats
The analytical process and simulation procedure for estimating the social potential is subject to various caveats. Most notably, realizing the social potentials shown in table 4 requires that all citizens have access to community investment options provided by reliable institutions.
This is not currently the case in all EU MSs.
Furthermore, the CE is designed to have changing profit rates according to different national market conditions, but the holding period is fixed for every country at 20 years. Switching to a non-fixed holding period for the simulation procedure would provide a more flexible maximization framework that would most likely result in a generalised increase of the estimated social potentials for every sampled country.
The same rationale could be employed when calibrating the 'RE technology' variable: based on the responses obtained, the analytical process takes the most preferred RE technology/installation on average across the EU-28, despite the potential for regional heterogeneity in this dimension. The proposed SBSS methodology allows for a more refined analysis by disaggregation in many dimensions, including in the dimension of the preferred RE technology on a country basis to determine whether some countries prefer solar or wind technologies. This might lead, again, to an overall increase of the estimated social potential (albeit only in countries with a strong preference for solar investments). The CE gave each respondent eight choice scenarios, but stipulated that the respondent consider them separately, not as additional investments to those already chosen in the exercise. For this reason, we only consider one RE technology (wind) as the object of investment. Simultaneously considering solar would violate this condition of the CE and in effect 'double-count' the willingness of respondents to participate in some CRE schemes.
In the same vein, the optimal capital requirement was calculated to be nation-specific, as our analysis focuses on the representative individual of each sampled country. However, future SBSS work could further disaggregate by income bracket, gender or age. In the present case this would result in a more appealing investment proposition being offered to different groups of citizens and would again likely lead to an overall increase in the estimated social potential.
All these measures, when combined, would likely increase individual respondents' investment probabilities, as well as the quantities willing to be invested. It would in turn increase countries' social potentials and, consequently, the GHG abatement potential of individual citizens across the EU. In this regard, the current analysis is considered a conservative estimation.
It is also important to acknowledge the effects that a hypothetical bias may have in respondents' manifested choices according to potential deviations from actual investment behaviours when confronted with a similar investment proposition in a real-life situation. This might lead to an over-estimation of the social potentials of each MS, and therefore inaccurately depict the real level of interest and willingness to invest of the average individual respondent.
In an attempt to account for the effect of hypothetical bias in the CE, the survey asked respondents if they would like to share their email to receive information on actual investment options from companies that offer CRE investments. This exercise exposed respondents to a small real-world 'cost' of sifting through future emails, and served to measure the sincerity of their interest in the CRE investments options. Almost half (48%) of CE participants explicitly stated their interest in receiving such information periodically and allowed access to their email addresses. Although the high proportion of respondents sharing their email suggests a sincere interest in real-world CRE schemes, we detected some responses that were at odds with choice behaviour. In particular, 1,963 respondents who chose to invest in all eight choice scenarios but then did not give their email for follow-up; and 697 respondents who chose not to invest in all scenarios but then gave their email. We consider these two groups of respondents to be candidates for hypothetical bias. As a robustness check, we drop all 2,660 respondents from the sample and re-run the SBSS procedure detailed in section 3.1. The full results of this exercise are not reported for brevity (available upon request), but the final EU-28 social potential is estimated at €151 billion when the potentially biased responses are removed. This is a 14% decrease from the original €176 billion estimated in table 4, yet would still represent a substantial contribution for financing next decade's €340 billion required investment in RE capacity.

Discussion
It is important to note that the experimental design of the research presented herein assumes that all citizens have access to market information and community investment options provided by reliable institutions. Most notably, it guarantees a risk-free investment operationalised through a trustworthy and straightforward financial vehicle. These facts, along with the availability of multiple investment options, offers a plausible explanation for the substantial social potential and high acceptance rates expressed for the investment offerings. Therefore, the results outlined above substantiate the need to ensure that easily accessible, trustworthy, and risk-insured community investment options are available across EU MSs to unlock their social potentials for investing in 25 CRE. This section presents policy-relevant insights that would help move towards a riskminimised regulatory environment for citizen-led RE finance and thus help satisfy the main policy recommendation that flows from our empirical analysis.

The effect of RE subsidies on the estimated social potential
In line with previous research on the role of financial participation in the social acceptance of RE [29], [31]- [33], the findings presented in this research suggest a tangible relationship between financial participation and co-ownership, and increased acceptance for localised forms of RE generation from European citizens.
It is important to highlight that the empirical analysis presented in this study estimates the EU's social potential for collectively investing in community-administered wind farm cooperatives under current market conditionsthat is, without capital subsidies or any other national support mechanism for RE development such as feed-in policies, investment/production tax credits, or other fiscal incentives. Our results therefore stem from a conservative estimation and reflect a subsidy-free social potential. However, national support schemes to RE generation are commonly employed throughout the EU-28. In order to understand the potential of subsidy schemes to influence citizens' willingness to co-invest in a community-administered wind farm development, we perform a scenario analysis using the SBSS process.
In order to do so, we take current national RE subsidies [57] and use these to re-calculate the profit rates for wind energy generation, as explained in section 3.1 (see Table A.3. in Appendix for details). We then re-conduct the SBSS analytical procedure to assess the changes in each country's average respondent's probability of accepting the most-preferred investment option incorporating a subsidised (as opposed to market-based) profit rate for their investment, and quantify the resulting social potential stemming from those probability changes (see Table A.4 in Appendix for details). The results stemming at the conclusion of this analytical procedure indicate that a subsidised social potential would yield a 27% increase in the expected volume of citizen investments collected across the EU, and reach a total volume of €224 billion ( figure 4). This equates to 1.3% of the EU's GDP in 2018. Thus, our social simulation results suggest that subsidies can play a role in increasing the total volume of CRE wind capacity realized. Though this result is still subject to the caveat discussed above, namely that citizens have access to, and awareness of, CRE investment options, policy strategies to make this a reality are discussed in the following subsections.

Regulatory risk, market instability and investor confidence
One possible reason for the high levels of interest observed in the CE may be the low-risk,  [2], [3], [60] and, as such, undermine the ability of European citizens to partake in CRE schemes and co-benefit from Europe's low-carbon energy transition.

Citizen agency in the energy transition
These insights attempt to contextualize this research into the ongoing transformation that energy markets must undergo for accommodating the transition to a low-carbon energy system with an increasingly diverse set of actors combining traditional players with newly emerging market participants. In that respect, the idea of social potential developed through this research may serve as a useful conceptual tool to further explore how the Energy Union's regulatory framework can be operationalised for supporting more inclusive and participatory pathways towards a decarbonised energy future. Collective finance for alternative energy generation schemes shaped by collaborative dynamics around local communities offers a vehicle of collective action leading to such citizen empowerment and the development of shared agency.
Extending the availability and awareness of CRE investment options could be a potentially resourceful approach in promoting the uptake of CRE schemes throughout Europe. Specifically, the provision of easily accessible, trustworthy, and low-risk community investment options and revenue mechanisms may be a viable means to expedite the pace of RES deployment and increase citizen participation through community-based forms of energy generation.

Conclusion
This study quantifies the social potential for participating in CRE investments to assess the feasibility of citizen-driven financing to bridge the next decade's investment gap for a decarbonised energy system. Using responses obtained from an international survey and choice experiment across all EU-28, a novel Survey-Based Social Simulation (SBSS) quantification method is developed and illustrated. The method relies on estimating the probability that the average representative citizen in each nation would participate in a CRE scheme with optimal investment and operational characteristics. These probabilities are imputed using a probabilistic discrete choice model anchored in economic random utility theory.
The results obtained indicate a substantial social potential of €176 billion that could be harnessed from European citizens willing to co-finance community-administered wind energy cooperatives with market-based rates of return. Realizing this social potential would be enough to halve the investment requirements foreseen to achieve a 32% RES share by 2030, leading to an aggregated energy generation potential of 195,805 GWh every year. This translates into an annual GHG emissions abatement potential of over 103 MtCO2-eq for the entire EU, equalling to a 2.3% annual reduction in EU-28 GHG emissions from 2018 levels. Introducing current RE subsidy schemes in the simulation procedure would result in a 27% increase on the estimated social potential and reach a total volume of €224 billion.
In light of the substantial interest and potential for participation in CRE initiatives, EU energy and climate policy must strive to generate trustworthy financial vehicles, stable regulation, and low-risk market conditions. Such a climate would facilitate the incorporation of more innovative yet risk exposed CRE developers (e.g. energy cooperatives) and help to expedite the increased penetration of RE in the EU.
While the SBSS method is subject to caveats and assumptions discussed above, the development and application of this method herein illustrates its potential for contributing to social research questions as a counterpoint to the rising popularity of agent-based modelling (ABM).
SBSS benefits from a standardized theoretical background (random utility theory) and data collection methodology (CE) vis a vis ABM. Whereas, ABM exhibits greater scope and flexibility in how research topics are addressed and how human behaviour is modelled. This flexibility also increases computational complexity especially of large-scale (e.g. EU-28) ABM endeavours that 29 can make the problems intractable [61]. For such large-scale problems social scientists may consider applying the SBSS method and corresponding simplifying assumptions.

Data Availability
The full dataset and code can be made available upon request or, alternatively, uploaded to the journal's archives.

Declaration of Interest
The authors declare no competing interests.