Master Thesis
Accelerating the energy
transition
A study on how governments can influence
households to invest in retrofit measures
Student: Fleur Mulder (S2519127)
Supervisor: Mr. Eelko Huizingh
Co-assessor: Mr. Machiel Mulder
Word count: 12111
MSc Business Administration – Strategic Innovation Management
University of Groningen
Current Dutch governmental polices on improving energy efficiency in houses show a variation in effectiveness. Combining existing studies on the effectiveness of governmental policies with existing literature on the effect of various factors influencing household’s
willingness to invest in retrofit measures leads to the development of a new framework. Based on the findings of this study, governments are provided with new theory on policymaking and guidelines to reach maximum effectiveness of policies aimed at stimulating energy-saving behavior. To test increased effectiveness of governmental policies, an experiment set-up is provided, based on social-influence based information campaigns, the theoretically most effective governmental policy.
Energy-saving, government, retrofit, households, social influence INTRODUCTION
prior diagnosis or evaluation, follow a scattergun approach or do not lead to ongoing activities (Gynther et al., 2011). Besides this, different households are motivated by different reasons, such as costs, perceived easiness or the environment. One approach to target all is not effective (Han et al., 2013). This study analyzes the efficacy of current energy efficiency policies and gives governments new insight in the possibilities to effectively influence households to make investments in retrofit measures.
Many papers have been written on the effect of socio-demographic, psychological, and home characteristics on energy-saving behavior and energy conservation (e.g. Frederiks et al., 2015). Literature on the effect of retrofit measure characteristics, such as financial costs and payback time, on energy-saving behavior exists in various forms (e.g. Schleich et al., 2017; Nair et al., 2010). Papers on the effect of social pressure on energy-saving behavior are available in high quantities (e.g. Nolan et al., 2008; Cialdini, 2007) Even literature on the effectiveness of current Dutch governmental policies to improve energy performance in houses is available (e.g. Murphy et al., 2012). However, to the knowledge of the writer, no literature is available on how governments can effectively influence households to make investments in retrofit measures by taking factors with a proven significant effect on energy-saving behavior into consideration. In other words, a literature gap exists on the combination of factors influencing energy-saving behavior and the effectiveness of current governmental policies. Can certain factors influencing energy-saving behavior improve the effectiveness of current – and future – governmental policies aimed at improving energy efficiency in houses?
This results in the following research question: “Which governmental measures influencing households to make investments in retrofit measures could be effective and how can the effectiveness of such measures be determined?”. In order to find an answer on this research question, four sub questions have been formulated:
Sub question 1: “Which factors influence households to invest in retrofit measures?” Sub question 2: “Which of these factors can be influenced by governments?”
Sub question 3: “By means of which policies can governments influence these factors?” Sub question 4: “Which fractions could occur between theory and practice while performing experiments to test the efficacy of policies influencing these factors?”
motivation to save energy, the role of households in the energy transition, and a summary of current policies performed by the Dutch government to motivate Dutch households to save energy. Moreover, the role of local governments and the difficulties they experience while carrying out national policies are being discussed. Afterwards, four factors possibly influencing household’s investment in retrofit measures by means of certain characteristics based on households, homes, and retrofit measures, and social influence are being discussed. Then, the possible influence of current governmental policies on these four factors is being analyzed, resulting in a theoretical framework, visualizing all possible influences. After a short first discussion of the collected literature, an experiment based upon the most interesting findings has been conducted. The set-up of the experiment is being discussed, as well as possible inconsistencies to arise between ideal and practice. Finally, this study aims to make a contribution to effective policymaking by the Dutch government to stimulate households to invest in retrofit measures.
METHODOLOGY
LITERATURE REVIEW Governments’ Motivation to Save Energy
At the Paris climate conference (COP21) in December 2015, 195 countries adopted a universal, legally binding global climate deal, setting out a global action plan to limit global warming to well below 2 °C in order to avoid dangerous climate change (European Commission, 2018). In order to limit total human-induced warming to less than 2°C, as proposed by the Paris climate conference, we can emit no more than about 800 billion tons of CO2. This is the so-called ‘carbon budget’ that we can emit in the future. Current emission levels lay around 40 billion tons CO2 per year, meaning that the remaining carbon budget will be entirely gone in 20 years (Peters, 2017).
At the current energy consumption rate, we might run out of fossil energy rather soon. The world population is provided with 115 remaining years of current coal production, and roughly 50 remaining years of both oil and natural gas if production were to continue at this rate (BP Statistical Review of World Energy, 2017). Taking this into consideration – fossil fuel reserves are limited, and the maximum accepted temperature will be reached in 20 years – it seems clear that a radical reduction in energy consumption and CO2 emission is needed.
consumption in 2050 (PBL, 2016). In order to reach the goals, set by the EU and the Dutch government, energy consumption and CO2 emission need to be reduced radically.
Governments on both European and national level are motivated to reduce their energy consumption for abovementioned reasons. For many national governments, including the Dutch, an additional motivation for reducing energy consumption is that saving energy helps to decrease a country’s dependency on foreign energy sources and keep energy costs affordable (House of Representatives, 2011).
In conclusion, governments are highly motivated to save energy to meet European and national energy-saving targets, postpone the bottom of the carbon budget, and become less dependent on other country’s fossil fuel reserves. In order to save energy, governments on all levels – European, national, and local – need to propose policies to reduce energy consumption of industry and households.
Role of households in saving energy
Total energy consumption in the Netherlands mounted to 3155 petajoule in 2016 (CBS, 2018a). The total energy consumption of Dutch households in that same year was 412,6 petajoule, which accounts for 13,1% of total energy consumption (CBS, 2018b). In terms of natural gas consumption, the main energy source to heat buildings, the percentage of household consumption compared to total consumption is 24% (CBS, 2018b). Whereas industry sectors need expensive, large scale solutions to decrease their energy and gas consumption, households can cut their energy consumption with relatively smaller investments and still induce a significant decrease in total Dutch energy consumption.
(Trotta, 2018). This study focusses on houses adopting energy efficient retrofit investments – as from now on referred to as retrofit measures.
Current Government Policies in the Netherlands
The Dutch Government currently performs several policies to influence households to save energy. The policies to be discussed in this section are specifically targeted at motivating households to invest in retrofit measures. Moreover, the efficacy of each policy is being discussed.
Subsidies and VAT reduction. The Dutch government provides subsidies to stimulate energy-saving in existing owner-occupied homes in the private sector, existing buildings owned by Associations of Owners, Housing Associations, and Housing Cooperatives (Blok, 2018). House owners and Associations can apply for the subsidy scheme Energy-Saving to fund various insulation measures. With the help of this subsidy, investors in insulation can retrieve 20% of the investment costs. However, this subsidy, like all national subsidies, is bounded by a budget. The budget for private home owners is through, meaning that they are not eligible to apply for this subsidy scheme anymore since April 2017. Associations of Owners do still have a claim to this subsidy until December 2018 (RVO, 2018).
The second subsidy currently available in the Netherlands is the subsidy for Sustainable Energy (ISDE). This subsidy provides monetary support for the investment in solar boilers, heat pumps, biomass heaters and pellet heaters and is available both private and business consumers (RVO, n.d.-1). Besides, owners of solar panels can reclaim the VAT they have paid over installed solar panels. In order to be eligible for this tax reduction, solar panel owners need to supply a certain amount of generated energy back to their energy supplier (Belastingdienst, n.d.).
In theory, subsidy-type measures are highly effective due to strong market sensitivity and a focus on the end-use sector with active stakeholder involvement (Lund, 2007). In the Netherlands subsides are not evaluated besides their user satisfaction level. Evaluation of Dutch subsidies concluded that 53% of Dutch subsidy receivers would have invested less – or not at all – in energy saving measures without a subsidy. The other 47% would have merely postponed investing in energy-saving measures or state that the subsidy had no influence on their decision to invest (Murphy et al., 2012). Thus, subsidies may be a critical factor for certain home owners to invest in energy-saving measures, while others are not affected by it.
more energy efficient house can opt for the Energiebespaarlening (Energy-saving loan). This loan can be used to invest in retrofit measures, such as insulation or a (solar)boiler. Solar panels are allowed to be financed by the loan as well, but for a restricted amount of 75% of the total loan. The remaining 25% needs to be invested in other energy efficiency measures (RVO, n.d.-2). Loans are theorized to be most helpful for low-income groups or starters on the housing market (Blom et al., 2009). It is thus that loans prove to be most effective for this particular group.
Energy Performance Certificate. Under the European Energy Performance of Building Directive, an Energy Performance Certificate (EPC) is required to sell or rent out a property. The EPC lists an energy rating for a building on a scale from A to G, where A is most energy efficient, and G the least (Murphy et al., 2012). Moreover, the certificate is issued by trained surveyors and contains specific advice on how to improve the energy efficiency of the building. It is assumed that this specific advice shows an economic benefit and stimulates consumers to purchase or rent a property (Murphy et al., 2012). EPCs are an effective policy measure, as they create a market demand for energy efficient houses. As of January 2017, over 3,2 million homes in the Netherlands have an EPC (CBS, 2017). Dutch properties with a label A, B, or C, are sold at an average price premium of 2,7% (Brounen & Kok, 2010). However, the proposed positive effect of information and advice disclosed in the energy labels on energy conservation is unreliable due to a lack of empirical testing (Murphy et al., 2012).
Building regulations. The national building decree sets requirements to the level of energy efficiency in new buildings according to the Energy Performance Coefficient (EPC) (RVO, n.d.-3). Certain requirements extend to home extensions and renovations and are expanded with renovation-specific minimum requirements for thermal resistance (Murphy et al., 2012). Regarding building regulations, the impact of energy efficiency requirements for renovations and extensions is minimal. Since building regulations for renovations and extensions strictly apply to the part of the house undergoing alternation, home owners are unwilling to extend energy efficiency measures to the whole property (Murphy et al., 2012). In order to stimulate sustainable renovation beyond the legal building regulations, the Dutch government offers a subsidy for sustainable property. The more energy efficiency measures are performed in the building, the higher the subsidy (RVO, n.d.-4). The effectiveness of the subsidy can be considered similar to the subsidies discussed earlier.
companies formally agreed to share responsibility to reach current climate change policy targets (Murphy et al., 2012). The main target of the covenant is to improve energy efficiency in 2.4 million buildings by 20-30% by 2020. Between 2008 and 2010, 20% additional energy savings were achieved in 314.000 properties. If home energy efficiency were to improve at this rate, it is unlikely that the 2020 target will be met (Murphy et al., 2012). Nevertheless, this policy measure can be considered a success since the policy has proven to effectively increase energy efficiency improvements in 314.000 houses.
Information campaigns. Information campaigns’ purpose is increase awareness and educate households about the energy efficiency possibilities in their homes. Information tools include generic policies, such as internet-based tools and TV broadcasts, and tailored advice, provided by certified companies (Murphy et al., 2012). The Dutch government offers a subsidy for energy advice, compensating for 75% of the invoice (RVO, n.d.-5). Information campaigns take time to be noticed and do not prove very effective in the first year after being set up. However, relevant organizations report an increase in enquiries over the past years in which information campaigns from the government were mentioned as initiator of interest in saving energy (Murphy et al., 2012). Tailored advice, combined with personal feedback, has proven to lead to a significant reduction in energy consumption and an increase in energy-saving behaviour. However, this effect was particularly found for relatively small behavioural changes and did not stress to larger investments requiring effort or money (Abrahamse et al., 2007). Measures on local scale
In all national energy policy plans, a large role has been assigned to municipalities. Local governments are in charge of the regional energy and heat plans to contribute to the national greenhouse gas emission reduction targets.
Besides, knowledge gaps obstruct municipalities to fulfil a directing role in the transition to sustainability. There is a need for information on how to organize a building team with specialist knowledge, knowledge about current policy instruments to stimulate adoption of energy efficiency investments (Tambach et al., 2010), and how to apply fitting sustainable alternatives to the right buildings and neighborhoods (Elzenga et al., 2017). Also, municipalities should get the jurisdiction to oblige property owners to switch to sustainable heating and electricity solutions, instead of replacing old nondurable installations with new nondurable ones (Elzenga et al., 2017).
In terms of municipalities’ ambitions, some local governments are more motivated to speed up the energy transition than others. Various incumbent municipalities are lacking motivation to renew old work patterns and work habits (Tambach et al., 2010). When the Dutch national government wants the urban environment to be greenhouse gas emissions free, binding agreements are required between local and national government (Elzenga et al, 2017).
Factors Influencing Investment by Households in Retrofit measures
Investments in retrofit measures by households are influenced by four factors, including; household characteristics, home characteristics, retrofit measure characteristics, and social pressure. Since these factors are collections of different sub characteristics, each with a unique (in)significant effect on investment in retrofit measures, these sub characteristics are to be discussed individually.
Household characteristics
Socio-demographic factors
Age. Age does not consistently emerge as a significant predictor of household energy use or energy-saving behavior. Whereas some literature expects older people to show more energy-saving behavior (Barr et al., 2005), other literature expects that middle-aged people tend to be the ones investing in retrofit measures most (Tonn & Berry, 1986). Altogether, prior research on the effect of age on investment in retrofit measures often results in inconsistent, minimal or statistically insignificant results (Frederiks et al., 2015). Therefore, age is not considered to have a significant effect on investment in retrofit measures.
Education. Education is associated with increased knowledge and awareness regarding energy-saving behavior but does not lead to this energy-energy-saving behavior per se (Kollmuss & Agyeman, 2002). It is not significantly proven that education has a major impact on performing energy-saving behavior (Frederiks et al., 2015). Therefore, in this study, education is considered to not have a significant effect on investment in retrofit measures.
Income. Household income is one of the strongest predictors of residential energy use, where higher-income households consume more energy than lower-income households (Frederiks et al., 2015). Where higher-income households use more energy, they also have an increased capability to invest in new appliances and retrofit measures. However, some studies suggest that not the highest-income households are most likely to save energy, but middle-income households are (Cunningham & Joseph, 1978, and Verhage, 1980). This conclusion stems from the finding that low-income households are not able to make investments in retrofit measures, while high-income households might not be interested in doing so. While some slight inconsistencies in the literature exist, a general conclusion can be made that a middle to high income does have a significant effect on investment in retrofit measures (Frederiks et al., 2015). Size. Households size is reported to have a positive impact on energy use as well, such that larger households consume in total more energy than smaller households (Frederiks et al., 2015). When energy consumption is measured per capita, however, single-person households use the highest amount of energy (Holloway & Bunker, 2006). Taking energy-saving into consideration, households consisting of three to four persons make greater investments in retrofit measures, report a greater history of past investments and expect less difficulty in making future investments (Curtin, 1976). Therefore, in this study, household size is considered to have a significant effect on investment on retrofit measures.
Psychological factors
this study does not consider personal values, attitudes & beliefs to have a significant effect on investment in retrofit measures.
Motives, intentions & goals. Personal motives, intentions & goals are expected to affect energy-saving behavior and investment in retrofit measures, but empirical proof is inconsistent. Moreover, intentions and plans alone do not necessarily lead to actual behavior and are only effective when combined with other moderating factors (e.g. socio-demographic factors or home characteristics) (Frederiks et al., 2015). Therefore, this study does not consider personal motives, intentions & goals to have a significant effect on investment in retrofit measures. Perceived responsibility. Feeling personally responsible for environmental problems affects energy-saving behavior, such that when people feel responsible, they feel more obliged to perform energy-saving behavior. However, general psychological literature often stresses the discrepancy between intentions and action, implying that the strength of this relationship may be weakened and unreliable due to the before-mentioned value-action gap (Frederiks et al., 2015).
Perceived behavioral control. People perceiving to have the capability to control events that impact them, i.e. a strong internal locus of control, are prone to be engaged in pro-environmental behavior, such as investment in retrofit measures (Frederiks et al., 2015). Consequently, individuals with an external locus of control, i.e. the belief that that decisions, circumstances and outcomes are controlled by environmental factors outside their influence, may be less likely to show energy-saving behavior because they perceive that such behavior is inefficacious (Kollmuss & Agyeman, 2002). As a result, this study considers perceived behavioral control to have a significant effect on investment in retrofit measures.
Risk aversion. Risk aversion is correlated with the adoption of retrofit measures, such that risk-averse households are less likely to invest in retrofit measures (Qiu et al., 2014) In other words, individuals who are more risk seeking, thus less risk averse, are more likely to renovate their homes and invest in retrofit measures (Fishbacher et al., 2015). However, the effect of risk aversion on investment in retrofit measures is found to be insignificant by others (Schleich et al., 2017). The effect of risk aversion on the adoption of retrofit measures is inconsistent, but the majority of literature available on the topic is in favor of a correlation. Therefore, this study joins the dominant literature stream and considers risk aversion to have a significant effect on investment in retrofit measure.
Home characteristics
free-standing or semi-detached homes typically consume more energy than households residing in smaller homes (Frederiks et al.,2015). Simultaneously, households living in larger homes are more willing to invest in retrofit measures as well. Some evidence indicates that a higher electricity or gas bill, due to more electronical appliances and a larger space to heat, might signal households that retrofit measures are desirable (Powers et al., 1992).
Age and condition. The age and condition of a house is positively related to energy-saving behavior as well, such that households living in older houses in a poor condition are more inclined to invest in retrofit measures in order to upgrade their homes (Nair et al., 2010). As a result, this study considers home’s age and condition to have a significant positive effect on investment in retrofit measures.
Ownership. Homeowners tend to be more likely to invest in large retrofit measures because they have greater financial security and receive a greater return on energy efficiency investments (Frederiks et al., 2015). Some research even states that home ownership is one of the most important factors affecting investment in retrofit measures (Black et al., 1985). This impact might be caused a sense of personal control and belonging, encouraging households to pay more attention to saving energy and thus making larger capital investments in energy-saving and retrofit measures (Barr et al.,2005).
Retrofit measure characteristics
Financial costs. In general, the price that homeowners are willing to pay for retrofit measures is higher than the actual costs of implementing these measures. However, the willingness to pay a certain price for retrofit measures is influenced by certain other factors, such as degree of concern for the environment (Banfi et al., 2005). While some people may be keen on optimizing their home energy efficiency, the immediate financial costs stemming from investment in retrofit measures might restrain less environmentally concerned people from doing so (Frederiks et al., 2015). Therefore, this study recognizes the effect of financial costs of retrofit measures on investment in these measures but assumes that this effect might be stronger for certain individuals than for others.
invest in energy efficiency retrofit measures of which rewards are only given in the future (e.g. reduced energy costs, environmental benefits), than in investments of which rewards are immediately received (Frederiks et al, 2015). The longer the payback period of an investment, the later rewards are received, the less people are inclined to make an investment. Retrofit measures, in particular solar panels, have high initial costs and a long payback period (Jafari & Valentin, 2017). Therefore, this study considers payback period to have a significant effect on investment in retrofit measures.
Future energy prices. During the lifespan of retrofit installations in houses, consumers need to reckon with an increasing energy price in the coming decades since a production maximum of mineral oil is to be expected (Jakob, 2006). Based on a future scenario in which energy prices will rise, the installment of various retrofitting measures is not expected to be profitable unless combined with a 50% reduction in power for lighting and appliances (Mata et al., 2011). However, individuals perceiving their household energy costs as high are more likely to invest in retrofit measures than individuals who perceive their energy costs to be low. Therefore, it is suggested that an increase in energy prices might motivate consumers to actively invest in retrofit measures and yield energy savings (Nair et al., 2010).
Comfort. Personal comfort, particularly the perceived loss of comfort resulting from investing in and installing of retrofit measures, has a strong effect on household energy usage. Moreover, when people perceive energy-saving behavior, including investment in retrofit measures, as leading to discomfort or a reduction in lifestyle quality, the likelihood of people engaging in energy conservation behavior decreases (Frederiks et al., 2015). Examining the kind of people willing to accept a certain loss in comfort (e.g. waiting longer for warm water or a lower house temperature) a distinction can be made between people highly concerned with the environment and people not concerned with the environment. Whereas 60% of people concerned with the environment are willing to sacrifice some comfort to save energy, only 25% of people not concerned with the environment are (Barr et al. 2005). Therefore, this study considers that perceived loss of comfort has a significant effect on investment in retrofit measures, but is moderated by other factors, such as environmental values.
Social influence
the adoption rate was higher than when anonymous or unpopular neighbors installed solar panels (Rogers, 1983).
Descriptive norms. Individuals are prone to imitate behavior that they observe to be done by other individuals (Melnyk et al., 2013). This effect can be explained by descriptive norms, which describe behavior of certain individuals and, whether consciously or unconsciously detected, show other individuals what is effective and desirable behavior (Melnyk et al., 2013). The same effect exists for pro-environmental behavior. Descriptive normative messages, describing energy-saving behavior performed by others, increases imitation of this behavior more than non-descriptive messages (Cialdini, 2007). Moreover, descriptive messages containing information on the energy conservation behavior of neighbors motivate consumers to save more energy than messages with standard appeals used to stimulate energy-saving, such as protecting the environment or saving money (Nolan et al., 2008). This study considers the use of descriptive norms promoting investment in retrofit measures to have a significant effect on investing behavior.
Reputation. Publicly disclosing a person’s environmental impact can provide motivation for energy conservation. This motivation stems from the wish to have a ‘green’ reputation and display behavior that is thought of as desirable by society. Individuals feel more social pressure when their energy-saving behavior, including prior investments in retrofit measures, has been made public. In order to reach their desired ‘green’ reputation and comply to society’s ‘green’ expectations, individuals are prone to change their behavior (Delmas & Lessem, 2014). Governmental Policies Influencing Households Characteristics, Home Characteristics, Retrofit measure Characteristics and Social Influence
The subsidies currently available in the Netherlands have a varying effectiveness. Where in 53% of the cases subsidies did influence the decision to invest in retrofit measures, in 47% of the cases it did not (Murphy et al., 2012). Unfortunately, to knowledge of the writer, no further research has been done on the explanation for this result. Analyzing why subsidies were not effective in 47% of the cases could provide an interesting insight in as how to adapt current subsidy policies to increase their effectiveness. However, considering that loans are theorized to be most effective for low-income groups (Blom et al., 2009), a link can be made. If the effectiveness of financial incentives is highest for low-income households, future policies should be aimed at this target group. The government cannot influence household characteristics, such as income, directly, but can take low-income households into consideration when implementing new subsidies.
Financial costs stemming from investment in retrofit measures might restrain individuals for whom these costs are too high (Frederiks et al., 2015). Moreover, people are less likely to invest in retrofit measures of which rewards are only given in the future. The longer the payback period, the less people are inclined to make an investment (Frederiks et al., 2015). Without a loan, the payback period – thus the time in which the consumer has fully recovered its costs – is long, since retrofit measures in nature have high initial costs (Jafari & Valentin, 2017). However, when making use of the Energiebespaarlening (energy-saving loan) made available by the government, the monthly costs of the loan can often be fully covered by the yield of installed retrofit measures (Nationaal Energiebespaarfonds, 2016). While in practice, consumers pay a higher total amount for the investment when making use of the Energiebespaarlening, they receive rewards from the first month onwards and are thus more inclined to invest in retrofit measures (Frederiks et al., 2015). Promoting the short-term rewards of the Energiebespaarlening more prominently might increase the effectiveness of this governmental policy.
thought of as desirable by society and the wish to gain a ‘green’ reputation (Delmas & Lessem, 2014). By taking the universal wish to receive a ‘green’ reputation into consideration governments could influence households to invest in retrofit measures to reach a ‘greener’ energy label.
Building regulations. Currently, building regulations for renovations strictly apply to the part of the house where the renovation takes place, leading to unwillingness of homeowners to extend installment of retrofit measures beyond the renovation (Murphy et al., 2012). Renovations usually take place in properties with a higher age and poor condition. It is suggested that households living in older houses with a poor condition are more inclined to invest in retrofit measures to upgrade their home (Nair et al., 2010). Therefore, it might increase building regulations’ effectiveness to target homeowners planning to renovate their house, thus homeowners with a relatively old house in a poor condition, with an information package about other retrofit measures that could be taken in their home.
More with Less covenant. The More with Less covenant is signed by the Dutch government, organizations representing the housing and building sectors, and energy companies (Murphy et al., 2012). However, the covenant could increase its efficacy when it includes individual home owners in the collaboration. By letting individuals signs the More with Less covenant, their perceived responsibility and perceived behavioral control might increase. These two psychological characteristics both have a significant effect on energy-saving behavior. When people feel responsible, they feel more obliged to perform energy-saving behavior (Frederiks et al., 2015). Moreover, when people perceive to have control over the behavior they feel responsible for, they are prone to be more engaged in pro-environmental behavior (Frederiks et al., 2015). Since currently organizations representing the housing sectors have signed the covenant, individuals might feel that these organizations take over responsibility and that they do not have direct control over the agreements in the covenant.
income and household size, since they have been proven to have a significant influence on investment in retrofit measures (Frederiks et al., 2015). In terms of psychological characteristics, perceived responsibility, perceived behavioral control, and risk aversion are the household characteristics to be included in consumer profiles. Regarding home characteristics, home type, size, age, condition, and ownership have a significant influence on investment in retrofit measures. In terms of retrofit measures, governments can target their information campaigns based on financial costs, payback period, future energy prices, and perceived loss of comfort. However, it has to be noted that the effect of financial costs on investment in retrofit measures is affected by other factors, such as concern for the environment (Banfi et al., 2005) and the height of immediate costs (Frederiks et al., 2015). When financial costs are taken into account in designing an information campaign, also several other characteristics from other factors need to be taken into account, such household income and perceived responsibility. By creating multiple consumer profiles based on abovementioned characteristics, governments make use of an approach called Database Marketing. This systematic and personal approach, specifically aimed to reach a certain target group, results in a more efficient use of information and communication resources and increases information campaigns’ effectiveness (Burghouts, 2013).
Theoretical Framework
Based on the literature on factors influencing energy-saving behavior or, more specifically, investment in retrofit measures, a theoretical framework can be constructed. In this framework, the influence of four factors (i.e. household characteristics, home characteristics, retrofit measure characteristics, social pressure) on investment in retrofit measures is shown. Additionally, the theoretical framework shows the indirect relationship of governmental policies on investment in retrofit measures via the four factors. Government policies can influence certain characteristics within each factor, which in their turn can promote investment in retrofit measures.
DISCUSSION
The four factors supposedly influencing investment in retrofit measures (i.e. household characteristics, home characteristics, retrofit measure characteristics, social pressure) have been analyzed by means of existing literature. Based on the findings, it is to be expected that certain sub characteristics of household characteristics, home characteristics, and retrofit measure characteristics do have a significant influence on investment in retrofit measures but cannot be directly influenced by governments. However, effective characteristics can be taken into consideration when conducting new policies. Where most current governmental policies, including subsidies, loans, EPC, building regulations and the More with Less covenant are targeted nationwide, it is expected that these policies might prove to become more effective when they would adopt a personal approach, targeted at a narrowly selected audience. Database Governmental
policies
Household
marketing would be an appropriate method to reach this personal approach. This approach is costly and time consuming, since information on certain household characteristics are difficult to gather (Burghouts, 2013). It is unlikely that small to medium sized governments, such as municipalities, would have the resources to operate this approach. However, since all current policies that could benefit from a personal approach are nationally initiated, it should be the national government to implement Database marketing on a nation-wide scale.
In terms of information campaigns, this policy would gain effectiveness by adopting a personal approach as well. Adapting current information campaigns to certain consumer profiles, based on various household characteristics and home characteristics, is expected to increase effectiveness of the policy. However, information campaigns could influence household’s investment in retrofit measures by means of a factor that the other current governmental policies cannot use: social influence. By making use of descriptive norms and taking the wish for a ‘green’ reputation into consideration, governmental information campaigns’ effectiveness could be drastically improved. Whereas the effect of social pressure on investment in retrofit measures has been confirmed (Wang, 2011; Rogers, 1983; Melnyk et al., 2013; Cialdini, 2007; Delmas & Lessem, 2014), the Dutch government has until now failed to fully incorporate social influence as a factor to motivate energy-saving behavior. Currently, the Dutch government plans to equip every household and small business with a smart meter by 2020 (Ministry of Economic Affairs, 2014). This meter provides households with their information on their energy consumption, and the energy consumption of an average Dutch household. At first glance, the smart meter could be an effective policy in which the effect of social influence in being incorporated. However, it is proven that the effect of social influence on pro-environmental behavior is stronger when consumers compared their own behavior with other consumers that are familiar, such as neighbors (Wang, 2011; Rogers, 1983; Nolan, 2008). Comparing households’ energy consumption with the consumption of an average, distant household, is t expected not to reach the save efficacy as comparisons with neighbors or other familiar persons.
Therefore, social influence provides opportunities for an effective information campaign, especially when social influence is being performed by consumers familiar to the targeted consumer. As a result, social influence might prove to be the base of an effective intervention used by governments to promote investment in retrofit measures.
Introduction
Governments could influence investment in retrofit measures by means of social influence-based information campaigns, as stated in the previous section. The level of social influence can be influenced by making use of individuals familiar to the targeted consumer, such that social influence may be higher when desirable behavior is performed by familiar individuals, than when desirable behavior is performed by unfamiliar individuals. In order to determine whether governments can influence the level of social pressure to such extent that it increases the propensity to invest in retrofit measures significantly, an experiment needs to be conducted. Experiments on the government’s influence on social pressure, indirectly influencing households’ investments in retrofit measures, have to the researcher’s knowledge not yet been conducted.
Hypotheses
Based on the literature review on social pressure performed in this paper, several hypotheses can be drawn.
Governmental information campaigns effectively influence households to invest in retrofit measures, be it after a certain time (Murphy et al., 2012). Governments are expected to influence household’s investment in retrofit measures when implementing a social influence-based policy that makes use of descriptive norms, individuals familiar to the targeted household, and the wish for a ‘green’ reputation., (Melnyk et al., 2013; Nolan, 2008; Delmas & Lessem, 2014).
H1: Governments can effectively influence households to invest in retrofit measures by means of policies based on social pressure.
Individuals are prone to behave the way they see others behave (Melnyk et al., 2013; Cialdini, 2007). Seeing is doing. Seeing neighbors invest in retrofit measures is thus expected to influence other households in the area to invest in retrofit measures as well. This results in hypothesis 2.
H2: Visibility of households’ investment in retrofit measures through placing a board on the house influences other households to perform the same behavior.
H3: Making an individual’s ‘good’ behavior, thus investing in retrofit measures, public affects other households’ investment in retrofit measures.
Experiment Set-up
This experiment will be conducted in the municipality of Groningen. The municipality will be divided in two parts, both being approached in a specific manner. In one of the parts, certain interventions will be put in place, while this will not be the case in the other part of the municipality. Over a period of time, the number of customer leads (customers asking for information), the number of offers for installations, and the number of installations performed will be gathered. The experiment’s purpose is to find out whether more customer leads, offers, and performed installations will occur in the part of the municipality where Reimarkt places interventions to stimulate social pressure, than in the part of the municipality where these interventions are not being placed. The collected data will be analyzed and will prove whether social pressure, when stimulated by the government, influences households’ investment in retrofit measures – or not.
In order to construct an experiment, conform to academic standards, the amount of rigidity, external factors influencing energy-saving behavior, and possible biases needs to be minimized. The experiment has been set up according to an ideal. Moreover, the experiment follows he protocol of Design of Experiments, which is defined as “the method of systematically obtaining and organizing knowledge so that it can be used to improve operations in the most efficient manner possible” (Condra, 2001, page 8). The method encourages to ask critical questions regarding every step and decision made in the experiment design. Such questions are ‘what could go wrong?’, ‘how will we know it went wrong?’, ‘what will we do if it goes wrong?’, and ‘how can we prevent it going wrong?’ (Condra, 2001). The experiment is designed according to a step-by-step procedure, where decisions have been made for every aspect. These decisions have been critically analyzed according to the Design of Experiments, thus defining which inconsistencies could arise between the ideal experiment set-up and practice. If such inconsistencies are expectable, actions to be taken in order to process the inconsistency or prevent it have been taken.
Step 2: Defining the target group. This experiment is conducted among households living in the municipality of Groningen. Moreover, participating households have to have an installation being executed in the time span of the experiment. Installations (investments in retrofit measures) include floor insulation, roof insulation, (cavity) wall insulation, glass insulation, heat pumps, and solar panels. Installations other than this provided enumeration or installations performed by organizations other than Reimarkt will not be part of the experiment.
Step 3: Mapping prior installations. The experiment will be executed in the municipality of Groningen. Based on prior installations performed by Reimarkt in 2017 and 2018, an estimate is made of the number of expected installations in the coming months. In the past 15 months (January 2017 till April 2018) 105 installations have been performed. An overview of the locations of these installations can be found in image 1 in the appendix. Of all installations, 18 are performed beyond the border of the municipality. These particular installations will be excluded from the estimation. Image 2 in the appendix shows the adapted visual. As a result, 87 installations have been performed by Reimarkt in the municipality of Groningen over the past 15 months. The installations have been performed gradually over time, resulting in 5,8 installations per month on average.
Inconsistency between ideal and practice. It may occur that during the time span of the experiment, significantly more installations are being performed in one part of the municipality than in the other. In theory, this chance is rather small, since the locations of previously performed installations have been taken into account in the division of the municipality in part A and part B. In practice, a difference in the number of initially planned installations may occur. This possible inconsistency between ideal and practice may have results for the collected data. When the number of installations performed in part A does not comply with control group B, a heightened amount of new installations cannot be concluded as being the result of social pressure-based interventions. The number and location of future locations cannot be influenced. When inconsistency between ideal and practice occurs, adaptations need to be made to the gathered dataset.
types of households and houses. Ideally, the experiment must be conducted amongst two completely equal groups, with an equal number of inhabitants and an equal mix of households and home characteristics in each part. Therefore, so-called neighborhood profiles have been created for every neighborhood in Groningen, based on certain socio-demographic characteristics, psychological characteristics, and home characteristics. The neighborhood profiles show house owner’s age, average household size, average income per working household member, average property worth, and percentage of privately owned homes. See table 1 in the appendix for a total overview of gathered data on neighborhood profiles. The decision to take these specific characteristics into consideration is based on the available data and the expected significant effect of these characteristics on energy-saving behavior (Frederiks et al. 2015). A visual division of the city into part A and part B can be found in image 3 in the appendix.
Inconsistency between ideal and practice. In practice, it is not possible to create an ideal division in which both parts of the municipality consist of perfectly equal neighborhoods. However, with help of the created neighborhood profiles, this division can be made approximately. Possible inconsistencies might arise during the experiment, when certain neighborhoods change in terms of household or home characteristics, resulting in a neighborhood that is more or less prone to invest in retrofit measures than was expected. Besides, in an ideal experiment, all characteristics expected to have a significant effect on energy-saving behavior would have been taken into consideration to form the neighborhood profiles. However, for certain characteristics expected to have an influence, data is unavailable. This is especially the case for psychological characteristics, which are difficult to measure. Step 5: Defining experiment’s physical scope. The experiment is conducted in the municipality of Groningen, not to be confused with the province of Groningen. The municipality of Groningen is limited to the city of Groningen, 11 villages, 2 former villages, 14 hamlets and 1 former hamlet (Plaatsnamengids, 2018). The province of Groningen consists of 20 municipalities, including the municipality of Groningen (Province of Groningen, 2018). A complete list of all neighborhoods participating in the experiment can be found in table 2 in the appendix.
at houses in these areas where installations were performed, people passing by might be influenced by the visual signs to invest in retrofit measures themselves. As a result, the influence of social pressure cannot be measured, since the people passing by do in all probability not know the people living in the houses where the installations take place. Therefore, the decision has been made to exclude certain neighborhoods and streets from the experiment. These excluded areas have been named buffer zones. Buffer zones are large ongoing roads, shopping streets, train stations, theatres, etc. A list of all neighborhoods in Groningen, divided into either part A or part B, and the corresponding buffer zones, can be found in the appendix as table 2.
However, buffer zones cannot eliminate all risks of people moving to places where they could be influenced in a way they are not supposed to. Ideally, people stay in their own neighborhood or at least stay within their part of the municipality in order to not get influenced by (the lack of) interventions placed in other neighborhoods. In practice, people visit family or friends in different parts of the city and see the interventions there. A person living in a neighborhood where interventions have not been installed, might consequently be influenced by interventions installed in other neighborhoods. If, as a result, this person decides to invest in retrofit measures, the data is rigged. It is, however, nearly impossible to find out if and why people are inspired to invest in retrofit measures, and whether they are inspired by signs they have seen in other parts of the city or not.
Step 6: Defining the starting point. The number of installations performed by Reimarkt is spread relatively equally over the year with 5,8 installations per month on average. The planning is to start the experiment in July 2018.
Inconsistency between ideal and practice. Ideally, the experiment could be performed year-round, without any implications. In practice, a slight decline in the number of installations can be expected during the Dutch annual closing for factories and construction companies. The annual closing for companies in the northern region – including Groningen – takes place from the 6th of August 2018 till the 24th of August 2018. This annual closing cannot be avoided, but to compensate for the possible reduction in performed installations, the experiment needs to be performed longer than originally planned. See step 7: Duration of the experiment.
Inconsistency between ideal and practice. Ideally, future customers stick to the average time needed to consider investing in retrofit measures. In that case, the first contracts for retrofit measure installations – influenced by the social pressure-based interventions - will be signed three months after the start of the experiment. In practice, customers often first search for information themselves on the internet, resulting in a longer process. Moreover, customers often go on holiday in the summer months, as has been introduced in Step 6. As a result, the process from information gathering to contract signing is prolonged with another three weeks – the duration of the Dutch annual closing. In order to provide customers with their needed time, the experiment’s time frame needs to be set up spaciously, requiring a minimum experiment duration of four months.
Step 8: Installation of interventions. The municipality of Groningen has been divided in part A and part B. In part A, a sign will be attached to the window of the house where the installation takes place, and cards will be distributed in the neighborhood. The sign will be attached to the front of the house and must be visible from the street. The cards will be posted in the mailboxes of 10 neighbors on each side of the house where the installation takes place, as well as 10 neighbors at the opposite side of the street. In case of a small street, with less than 30 neighbors, the cards will only be posted in the available neighbors’ mailboxes. In part B, no signs will be attached to the house performing the installation and no cards will be posted. This group is the control group.
Inconsistency between ideal and practice. Ideally, all participating households are willing to let a sign be placed on their house. In practice, house owners of a house where an installation takes place, located in part A, might refuse to have a sign being placed on their window. House owners may refuse the signs because of aesthetic reasons, possible expected damage to the window, or because of the time that the sign remains attached. In all of these cases, the data derived from that (part of the) neighborhood can in all expectancy not be used in the results of the experiment anymore. It is, however, important that the data is still collected, but is being accompanied with a note.
Inconsistency between ideal and practice. Ideally, the signs should remain attached to the house for a period as long as possible, to reach an audience as large as possible. In practice, it can be expected that most households will not agree with having a sign attached to their house for the total four-month duration of the experiment. Two weeks would be an appropriate compromise.
Step 10: Material content. In this experiment, signs to be placed on a window and cards to be posted in the neighbor’s mailboxes are used. The signs express that the house on which they are placed is being made sustainable by means of installing a certain retrofit measure. Moreover, the sign provides the viewer with contact details of Reimarkt and the encouragement to contact Reimarkt to ask for the possibilities for home improvement in the viewer’s home The cards express the address of the house in which the installation is performed, as well as the type of installation. The cards also contain the contact details and the encouragement to contact Reimarkt to ask for the possibilities for home improvement in the reader’s home. Moreover, a measurable URL is printed on both the signs and the cards. This link to a certain website (www.groningenwoontslim.nl/bord for the signs, and www.groningenwoontslim.nl/kaart for the cards) measures the amount of times that the particular URL is being visited. This data can be used to analyze whether there is a difference between the number of hits for the link on the signs and the link on the cards. The derived data can provide an insight in the effect of signs and cards.
Step 11: Material quantity. Reimarkt possesses 4 window signs and 400 cards with a measurable URL. For the experiment, this number of signs is sufficient. At an expected number of 5,8 installations per month, a sign will be placed in 50 percent of the cases, resulting in the rounded number of 3 required signs per month. One sign can be used as a buffer, in scenarios where an above average amount of installations takes place. 400 cards are sufficient for distribution at 13 installations, based on 30 required cards per installation. In the total time span of the experiment, 23,2 installations are to be expected. In 50 percent of the cases cards will be posted, resulting in around 12 situations where cards will be posted. The current number of cards is thus sufficient.
Step 12: Data Collection. Data will be collected by Reimarkt/Groningen woont Slim in a computer system called Pipedrive. All e-mails, telephone calls and other activities are registered in this system by the employees working at Groningen woont Slim. The gathered dataset will be made available to the University of Groningen and the responsible researcher after the experiment. The employees of Groningen woont Slim know how to operate the system and are aware of the exact data collection procedure (i.e. which data, when, how). The data to be collected includes the number of people that 1) requests information, 2) requests an offer, and 3) signs an offer. Customers requiring information can do so by calling to Groningen woont Slim by telephone, sending an e-mail, or visit the Groningen woont Slim shop. In the case that one customer approaches Groningen woont Slim in several ways, this will be registered as merely one entry. It is important to know where the customer lives, in order to know if their interest might be heightened by an intervention placed by Reimarkt, or not. Moreover, employees of Groningen woont Slim providing customers with information need to ask the customer how they have come to know about Reimarkt’s retrofit measures. During telephone calls and face-to-face conversations this question can be easily asked. However, when a customer requests information by e-mail, this e-mail is often replied to with standardized information. At a later point in time, customers are contacted by phone to ask whether all sent information was clear. At this moment, the employee of Groningen woont Slim is in the position to ask the customer how they have come to know Reimarkt.
these residents’ associations has not been taken into consideration in this experiment, although it is to be expected that their influence in practice is significantly high. This influence can lead to problems for the results of this experiment when, for example, an association is located in part A and it cannot be said with certainty that an increase in the customer leads from that particular neighborhood is caused by Reimarkt’s interventions or spiked by the association’s influence.
DISCUSSION & CONCLUSION
The main finding of this study suggests that the most effective way for governments to influence households to invest in retrofit measures, is by means of social influence-based information campaigns, which effectiveness could be determined by means of a practical experiment. In this study, the effectiveness of household characteristics, home characteristics, retrofit measure characteristics and social influence on households’ willingness to invest in retrofit measures has been analyzed. The findings are that in terms of household characteristics, household income and household size have proven to be effective socio-demographic characteristics influencing investment in retrofit measures, whereas perceived responsibility, perceived behavioral control, and risk aversion have proven to be effective psychological characteristics. Moreover, in terms of home characteristics, house type, size, age, condition, and ownership proved to have a significant effect on investment in retrofit measures. In terms of retrofit measure characteristics, the financial costs of the measure, payback period, future energy prices and perceived loss of comfort all effectively influence households’ willingness to invest in retrofit measures. Lastly, the findings suggest that social influence has a significant influence on investment in retrofit measures when opinion leaders, descriptive norms, and the wish for a ‘green’ reputation are made use of.
with a ‘green’ label are being sold at a premium. Our findings suggest that the effectiveness of EPCs can be increased by specifically promoting ‘green’ houses to provide a ‘green’ reputation to households, since this characteristic of social pressure has proven to be a significant determinant to show energy-saving behavior. In terms of building regulations, current effectiveness of this policy is low. The findings suggest that effectiveness could be increased when this policy incorporates the given that inhabitants of old houses, in need of renovation, are more willing to invest in retrofit measures, by means of providing an information package on possible other home improvements, for instance. The More with Less covenant is proven to be effective, but its effectiveness could be enhanced by involving individuals directly, instead of through housing associations. By doing so, perceived responsibility and perceived behavioral control would be enhanced, which both are significantly effective psychological characteristics influencing investment in retrofit measures. Information campaigns have not found to be effective immediately but can be on the longer term. The findings suggest that, by making use of database marketing, thus targeting households personally, and personalized normative feedback, showing households what individuals in their surroundings do, the effectiveness of information campaigns can improve tremendously, while making use of social influence. Since the Dutch government does not yet incorporate all possibilities of social influence in their information campaigns, but social influence might prove to be the base of an effective intervention to promote investment in retrofit measures, a possible experiment has been set up to test this theory. Experiments similar to these, where the effect of social influence through making desirable behavior by neighbors visible, is hypothesized to influence individuals to invest in retrofit measures significantly more than where desirable behavior of neighbors has not been made visible. However, since experiments in which human beings are involved can never be perfectly isolated from all factors possibly causing ‘noise’, possible inconsistencies between the ideal experiment and practice have to be taken into consideration.
households’ willingness to invest in retrofit measures, other than social influence, could emerge. If so, this factor would need extensive research, before a new experiment can be conducted, taking the newly found factor into consideration.
This research contributes to literate on energy-saving behavior, by closing the research gap of how governments can effectively influence households to make investments in retrofit measures by taking factors with a proven significant effect on energy-saving behavior into consideration. Furthermore, this research practically contributes to a more successful and streamlined energy transition by providing governments with a thorough advice in which knowledge from multiple literature fields – various factors influencing households’ willingness to invest in retrofit measures, and the effectiveness of current governmental policies – is combined into one framework. By specifically analyzing how current governmental policies could be improved by means of household characteristics, home characteristics, retrofit measure characteristics, and social influence, the findings of this study suggest that especially social influence-based information campaigns prove to be an effective policy for governments aiming to promote energy-saving behavior amongst households.
REFERENCES
Abrahamse, W., Steg, L., Vlek, C., & Rothengatter, T. (2007). The effect of tailored information, goal setting, and tailored feedback on household energy use, energy-related behaviors, and behavioral antecedents. Journal of environmental psychology, 27(4), 265-276. https://www-sciencedirect-
com.proxy-ub.rug.nl/science/article/pii/S0272494407000540?_rdoc=1&_fmt=high&_origin=gateway&_docanch or=&md5=b8429449ccfc9c30159a5f9aeaa92ffb&ccp=y#bib36
AlleCijfers.nl (2018). Overzicht van de 124 wijken en buurten in Groningen. Accessed on 23.05.2018. Retrieved from: https://allecijfers.nl/gemeente-overzicht/groningen/
Banfi, S., Farsi, M., Filippini, M., & Jakob, M. (2008). Willingness to pay for energy-saving measures in residential buildings. Energy economics, 30(2), 503-516.
https://www.ethz.ch/content/dam/ethz/special-interest/mtec/cepe/cepe-dam/documents/research/cepe-wp/CEPE_WP41.pdf
https://www-sciencedirect-com.proxy-ub.rug.nl/science/article/pii/S0301421503003859?_rdoc=1&_fmt=high&_origin=gateway&_docanch or=&md5=b8429449ccfc9c30159a5f9aeaa92ffb
Belastingdienst. (n.d.). Eigenaren van zonnepanelen. Accessed on 17.08.2018. Retreived from:
https://www.belastingdienst.nl/wps/wcm/connect/bldcontentnl/belastingdienst/zakelijk/btw/hoe_werkt _de_btw/voor_wie_geldt_de_btw/eigenaren-van-zonnepanelen/eigenaren_van_zonnepanelen
Black, J.S.; Stern, P.C.; Elworth, J.T. (1985). Personal and contextual influences on household energy adaptations. J. Appl. Psychol. 70, 3–21. Accessed on 15.06.2018. Retrieved from:
http://psycnet.apa.org/record/1985-16122-001
Blom, M. J., Korteland, M. H., & Schepers, B. L. (2009). Effecten en uitwerking van een Energiebesparingsfonds. CE Delft, Delft. Accessed on 18.06.2018. Retreived from:
https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=2&cad=rja&uact=8&ved=0ah UKEwiC9d7Xve7bAhWM_KQKHVEEDswQFggsMAE&url=https%3A%2F%2Fwww.cedelft.eu%2 Fen%2Fpublications%2Fdownload%2F778&usg=AOvVaw2rHwR422M4Yc8c9lXCa5KG
Blok, S. A. (2018). Subsidieregeling energiebesparing eigen huis. Accessed on 18.06.2018. Retreived from: http://wetten.overheid.nl/BWBR0038472/2018-04-04
BP. (June 2017). BP Statistical Review of World Energy 2017. Retrieved January 22, 2018 from:
https://www.bp.com/content/dam/bp/en/corporate/pdf/energy-economics/statistical-review-2017/bp-statistical-review-of-world-energy-2017-full-report.pdf
Brounen, D., & Kok, N. (2011). On the economics of energy labels in the housing market. Journal of
Environmental Economics and Management, 62(2), 166-179. Accessed on 18.06.2018. Retrieved
from:
https://www-sciencedirect-com.proxy-ub.rug.nl/science/article/pii/S0095069611000337?_rdoc=1&_fmt=high&_origin=gateway&_docanch or=&md5=b8429449ccfc9c30159a5f9aeaa92ffb
Burghouts, H., de Kleijn, B., van Leerdam, W. (2013). Databasemarketing voor energiebesparing: Praktijk gericht onderzoek. Ministry of the Interior and Kingdom Relations. Accessed on 12.06.2018. Retreived from:
CBS. (2017). Energielabels van woningen, 2007 – 2016. Centraal Bureau voor de Statistiek (CBS),
Den Haag; PBL Planbureau voor de Leefomgeving, Den Haag; RIVM Rijksinstituut voor Volksgezondheid en Milieu, Bilthoven; en Wageningen University and Research, Wageningen.
Accessed on 22.06.2018. Retreived from: http://www.clo.nl/indicatoren/nl0556-energielabels-woningen?i=16-53
CBS. (2018a). Energieverbruik per energiedrager, 1990-2016. Centraal Bureau voor de Statistiek
(CBS), Den Haag; PBL Planbureau voor de Leefomgeving, Den Haag; RIVM Rijksinstituut voor Volksgezondheid en Milieu, Bilthoven; en Wageningen University and Research, Wageningen.
Accessed on 19.06.2018. Retreived from: http://www.clo.nl/indicatoren/nl0054-energieverbruik-per-energiedrager-
CBS. (2018b). Energieverbruik door huishoudens, 1990-2016. Centraal Bureau voor de Statistiek
(CBS), Den Haag; PBL Planbureau voor de Leefomgeving, Den Haag; RIVM Rijksinstituut voor Volksgezondheid en Milieu, Bilthoven; en Wageningen University and Research, Wageningen.
Accessed on 19.06.2018. Retreived from: http://www.clo.nl/indicatoren/nl0035-energieverbruik-door-de-huishoudens
Centraal Bureau Statistiek Statline .(2018). Kerncijfers wijken en buurten 2017. Accessed on 22.05.2018. Retrieved from:
http://statline.cbs.nl/Statweb/publication/?DM=SLNL&PA=83765NED&D1=0-1,3-4,14,31,34,39,73,77,80,90,99,103-104&D2=9821-9918&VW=T
Cialdini, R. B. (2007). Descriptive social norms as underappreciated sources of social control. Psychometrika, 72(2), 263. Accessed on 05.06.2018. Retrieved from:
https://link.springer.com/article/10.1007/s11336-006-1560-6
Condra, L. (2001). Reliability improvement with design of experiment. CRC Press. Accessed on 11.06.2018. Retrieved from:
https://books.google.de/books?hl=nl&lr=&id=xuda-kq90OIC&oi=fnd&pg=PR5&dq=design+of+experiment&ots=QjkEvF6qvq&sig=QUEralrJmfzgEr5o RnIX-575Viw#v=onepage&q=design%20of%20experiment&f=true
Crosbie, T. (2006). Household energy studies: the gap between theory and method. Energy &
Environment, 17(5), 735-753. Accessed on 05.06.2018. Retreived from:
Cunningham, W.H.; Joseph, B. (1978). Energy conservation, price increases and payback periods.
Association for Consumer Research: Ann Abor, MI, USA. Volume 5, pp. 201–205. Accessed on
16.06.2018. Retrieved from: http://www.acrwebsite.org/volumes/9424/volumes/v05/NA-05 Curtin, R.T. (1976). Consumer adaptation to energy shortages. J. Energy Dev. 2, 38–59. Cuug. (n.d.). Energy and Power. Accessed on 22.06.2018. Retrieved from:
http://www.cuug.ab.ca/branderr/nuclear/petajoule.html
Delmas, M. A., & Lessem, N. (2014). Saving power to conserve your reputation? The effectiveness of private versus public information. Journal of Environmental Economics and Management, 67(3), 353-370. Accessed on 04.06.2018. Retrieved from:
https://www.sciencedirect.com/science/article/pii/S0095069614000072
Elzenga, H., Schwencke, A. M., van Hoom, A. (2017). Het handelingsperspectief van gemeenten in de energietransitie naar een duurzame warmte- en elektriciteitsvoorziening. Den Haag: PBL. Accessed on 18.06.2018. Retreived from:
http://www.pbl.nl/sites/default/files/cms/publicaties/pbl-2017-het-handelingsperspectief-van-gemeenten-in-de-energietransitie-1955.pdf
Enerdata. (2017). Global Energy Statistical Yearbook 2017: Breakdown by country (Mtoe). Accessed on 18.06.2018. Retreived from: https://yearbook.enerdata.net/total-energy/world-consumption-statistics.html
Epper, T., Fehr-Duda, H., & Schubert, R. (2011). Energy-using durables: the role of time discounting
in investment decisions (No. 11-16). IED Institute for Environmental Decisions, ETH Zurich.
Accessed on 12.06.2018. Retrieved from: https://www.thomasepper.com/papers/TRIEReport.pdf European Commission. (updated 2018). Paris Agreement. Accessed on 22.01.2018. Retreived from: https://ec.europa.eu/clima/policies/international/negotiations/paris_en
European Commission. (updated 2018). 2050 low-carbon economy. Accessed on 22.01.2018. Retreived from: https://ec.europa.eu/clima/policies/strategies/2050_en
Fischbacher, U., Schudy, S., & Teyssier, S. (2015). Heterogeneous preferences and investments in
energy saving measures (No. 2015-11). Munich Discussion Paper. Accessed on 12.06.2018. Retrieved
