Introduction
A significant share of energy production in the United States, around 80%, comes from fossil fuels (Sanchez Reference Sanchez2020). This presents a major challenge to reduce CO2 emissions under the Affordable Clean Energy (ACE) rule. While coal consumption has declined under ACE, it still remains a major source of energy production, resulting in local emissions and greenhouse gases (EIA 2020). The most effective method to reduce emissions and greenhouse gases is by shifting toward renewable energy sources, such as biomass, solar, and wind sources (IEA Reference IEA2017). However, because the cost of producing energy from these new sources under current technology is higher than from fossil fuels, this shift largely depends on households’ willingness to pay (WTP) for renewable energy, which is driven by their prosocial desire. Households’ prosocial behavior is the intrinsic motivation to take action in a community’s best interest (Bénabou and Tirole Reference Bénabou and Tirole2006). Previous research indicates that households’ willingness to be taxed serves as a signal of their prosocial identity, leading to subsequent behavior in line with that self-perception (Gneezy et al. Reference Gneezy, Imas, Brown, Nelson and Norton2012). When households recognize that their contributions help society, there is an increased likelihood of engaging in prosocial action and demonstrating prosocial WTP (Guerra and Harrington Reference Guerra and Harrington2021; Thornton et al. Reference Thornton, Aknin, Branscombe and Helliweal2019).
Households are increasingly aware and concerned about the production process of products and services, and are now valuing it in addition to the finished products. Recent literature reveals that consumers now prefer products that exhibit eco-friendly traits, such as animal welfare and fair-trade practices (Loureiro and Lotade Reference Loureiro and Lotade2005; Lai et al. Reference Lai, Wang, Ortega and Widmar2018; Delmas and Gergaud Reference Delmas and Gergaud2021). In the context of power generation, households’ prosocial behavior/green pricing, measured by their WTP for renewable energy, may vary depending on the type of renewable energy source, such as (biomass, solar, and wind.Footnote 1 Existing literature does not address the possibility of prosocial behavior in power generation. The objective of this study is to fill this knowledge gap by examining the extent of prosocial behavior/green pricing for renewable energy sources. We hypothesize that the WTP for renewable energy sources will be positive but will vary across different types of renewables and household prosocial characteristics, especially attitude toward the environment.
We test the hypothesis using choice experiments conducted in the United States. Households are presented with a choice of energy produced with different percentages of renewable sources blended with fossil fuel at alternative price levels. Results from choice experiments are employed in the context of random utility theory to estimate households’ marginal WTP for competing energy sources.
Our choice experiment and WTP analysis contribute to the knowledge base of household values for alternative renewable energy sources. We build on the work of Noblet et al. (Reference Noblet, Teisl, Evans, Anderson, McCoy and Cervone2015), which focused on the choices households made when asked to express their willingness to support proposed energy policies. Our extension provides estimates on the magnitude of households, by region and demographics, on WTP for alternative renewables (biomass, solar, and wind). Our research design improves previous research and contributes to the literature in three unique ways. First, we differentiate between three sources of renewable energy sources (biomass, solar, and wind) as compared to a single renewable source or an aggregate of all renewable sources as done in previous literature. Second, the experimental design incorporated individual households’ base energy consumption. This allows us to anchor price premiums and avoid overestimation of WTP. Third, we also incorporate the effect of emotions related to climate change in estimating households’ WTP for renewables. Specifically, our results indicate that households’ WTP for renewables varies by type of energy source. Prosocial behavior, as indicated by WTP for renewables, is positive for solar power, whereas no prosocial behavior toward their adoption of biomass and wind power is observed. Although our empirical test confirms our hypothesis of varying WTP across renewable energy sources, the nonpositive WTP for biomass and wind deserves some attention.
Our results are not restricted to just an empirical WTP exercise. They have major policy implications. In August 2022, Congress passed an extension of the Solar Investment Tax Credit, raising it to 30% for solar installation. The question is whether such a subsidy was required given the relatively high level of prosocial behavior existing with solar technology. This is in contrast to biomass and wind, which do not exhibit such high prosocial choices. Failure to consider this embedded behavior can lead to inefficient policies at best and failed policies at worst. In the following, we will present the literature review first, then explain our empirical framework, followed by the experiment design. Results, policy discussion and conclusions are presented at the end.
Literature
There exists literature assessing the WTP for renewable energy sources focusing on the environmental attributes associated with them. For instance, some studies have estimated the WTP for utility investments in renewable power generation (Zarnikau Reference Zarnikau2003; Bergmann et al. Reference Bergmann, Hanley and Wright2006; Scarpa and Willis Reference Scarpa and Willis2010), while others have estimated the WTP for increased renewable energy (Byrnes et al. Reference Byrnes, Miller and Schafer1999; Ethier et al. 2000; Gossling et al. Reference Gossling, Kunkel, Schumacher, Heck, Birkemeyer, Froese, Naber and Schliermann2005; Borchers et al. Reference Borchers, Duke and Parsons2007; Longo et al. Reference Longo, Markandya and Petrucci2008; Li et al. Reference Li, Jenkins-Smith, Silva, Berrens and Herron2009; Erdem et al. 2010; Vainio et al. Reference Vainio, Paloniemi and Varho2017; Bamwesigye Reference Bamwesigye2023). However, these studies are limited in their scopes as they focus on only a few specific renewable energy sources or specific regions. For example, Roe et al. (Reference Roe, Teisl, Levy and Russell2001) estimated the WTP for renewable energy’s environmental attributes without delving into individual preferences for renewable energy sources. Similarly, Borchers et al. (Reference Borchers, Duke and Parsons2007) compared WTP for renewable energy by the source to generic “green energy” in Delaware., while Mozumder et al. (Reference Mozumder, Vásquez and Marathe2011) investigated households’ WTP for a renewable energy program in New Mexico. Murakami et al. (Reference Murakami, Ida, Tanaka and Friedman2015) performed a comparative analysis of WTP for nuclear and renewable energy between four states within the United States and Japanese households.
Several studies have elicited consumers’ WTP for renewable energy using methods, such as contingent valuation surveys (Koundouri et al. Reference Koundouri, Kountouris and Remoundou2009; Zografakis et al. Reference Zografakis, Sifaki, Pagalou, Nikitaki, Psarakis and Tsagarakis2010; Mozumder et al. Reference Mozumder, Vásquez and Marathe2011) and discrete choice experiments (Roe et al. Reference Roe, Teisl, Levy and Russell2001; Scarpa and Willis Reference Scarpa and Willis2010). Discrete choice experiments are effective in eliciting consumers’ WTP based on the energy bundles’ attributes (Boxall et al. Reference Boxall, Adamowicz, Swait, Williams and Louviere1996). The participant’s choice reflects the tradeoff they make among the attributes. In our case, the product is the energy bundle, and the attributes are the renewable energy sources blended in the bundle.
Previous research did not consider consumers’ energy bills before presenting price increases to them (Borchers et al. Reference Borchers, Duke and Parsons2007; Mozumdar et al. Reference Mozumder, Vásquez and Marathe2011; Murakami et al. Reference Murakami, Ida, Tanaka and Friedman2015). Consumers were only asked to consider the dollar increase to their monthly bills, without considering the percentage increase of the renewable energy component nor average energy expenditures (Longo et al. Reference Longo, Markandya and Petrucci2008). Longo et al. (Reference Longo, Markandya and Petrucci2008) estimated a 54% price premium for renewable energy, which is larger than one standard deviation in energy expenditure. With price premiums not anchored in energy bills as a percentage, it may lead to an overestimation of WTP, the reason why some studies employ percentage price changes such as Ortega et al. (Reference Ortega, Wang, Widmar and Wu2014). It is important to anchor price premiums from a policy design perspective. Policy decisions are usually made as a percentage increase rather than large lump-sum increases on bills regardless of usage (United States Energy Information Administration 2009).
Previous studies have shown that negative emotions, such as worry and fear, have a stronger impact on national policy support compared to positive emotions (Smith and Leiserowitz Reference Smith and Leiserowitz2014; Rodriguez-Sanchez et al. Reference Rodriguez-Sanchez, Schuitema, Claudy and Sancho-Esper2018). Smith and Leiserowitz (Reference Smith and Leiserowitz2014) found hope to be a positive predictor of climate change policy acceptance, while Truelove (Reference Truelove2012) discovered that positive emotions were associated with support for wind energy, and coal and nuclear were viewed more negatively. On the other hand, a study by Sjöberg (Reference Sjöberg2007) revealed that negative emotions, such as fear and anger can influence the perception of genetically modified food, cell phones, terrorism, and radiation. Consumers’ valuation represented by their WTP has also been found to be influenced by their emotions. For example, Kessler et al (Reference Kessler, McClellan, Nesbit and Schotter2022) found short-term emotions influence economic decision-making.
Built upon the strands of these studies, we employ choice experiments to elicit consumers’ WTP for alternative renewable energy sources, considering their emotions. The WTPs are measured by percentage increases in consumers’ energy bills when their energy mix has one percent increase in a particular renewable energy source. We also employ a national survey representing consumers in the entire United State.
Empirical framework
In contrast to treating commodities as homogeneous in quality and divisible in quantity, recent studies focus on product quality attributes. Lancaster utility theory (Reference Lancaster1966) is the paradigm, that assumes households derive utility from attributes of commodities they consume. Households’ decision is then what commodity attributes they would select instead of quantities to maximize utility. This is a hedonic study (Rosen Reference Rosen1974) and describes our study of energy type selection.
Random utility models, developed by McFadden (Reference McFadden1975), are generally adopted for empirical studies employing survey data measuring consumer-stated preference. The randomness in utility is a result of either the random preferences of survey participants or a random subset of the complete set of information available to participants. In this framework, consider a household exposed to n combinations of alternative energy sources, then the random utility function for household i with combination j is
$${U_{ij}} = {V_{ij}} + {\varepsilon _{ij}},\; \quad j= 1\;{\rm{to\;}}n,\;$$
where
${V_{ij}}$
is the non-stochastic utility function and
${\varepsilon _{ij}}$
is the stochastic component.
We consider four energy sources (biomass, solar, wind, and fossil), with denoting the share of source l in the energy mix. For the traditional fossil source (coal, natural gas, or oil), l = 0 and l = 1, 2, and 3 represent wind, solar, and biomass, respectively. Thenon-stochastic part is then
$${V_{ij}} = \;{\beta _{iP}}{P_{ij}} + {\Sigma _l}{\beta _{il}}{x_{ijl}} + {\beta _{i0}},$$
where
${P_{ij}}$
is household i’s price premium for the jth combination of alternative energy sources, and
${x_{ijl}}$
denotes for household i’s share of energy source l in combination j. Variable
${x_{ijl}}$
is the quantitatively measured attribute, which contributes to utility (see detailed description in the choice experiment design section below). With all shares adding up to one, we normalize the traditional energy
${x_{ij0}}$
to be one for all i and j. Coefficient
${\beta _{iP}}$
is the marginal utility of price for household i which should be negative, and
${\beta _{il}}$
for l = 0, 1, 2, and 3 is the marginal utility of each energy source.
Household i’s marginal WTP for energy source l is the utility tradeoff, where it is indifferent between paying a price premium for consuming a one percent increase in energy l or not
$$WT{P_{il}} = - {{\partial {V_{ij}}} \over {\partial {x_{ijl}}}}/{{\partial {V_{ij}}} \over {\partial {P_{iP}}}} = - {\beta _{il}}/{\beta _{iP}}.$$
In a choice experiment, each household is presented with m choice sets with each set containing the three energy sources (biomass, solar, and wind). Each household is asked to choose a single source k from each choice set. Household, i, will choose k only when it provides the highest utility among the three. Denoting Y as the observed choice number
$$Prob{\rm{ }}\left( {{Y_i}\; = \;k} \right)\; = \;Prob{\rm{ }}({U_{ik}} \ge {U_{il}}),\ for\ \ l \ne k,\,{\rm{and}}\,\,l\; = \;1,{\rm{ }}2,{\rm{ }}3.$$
A random parameter logit model is employed for estimating the attribute coefficients in (2) using a choice experiment survey with n participants, m choice sets, and s sources (total mns observations). Allowing for household heterogeneity, the estimated coefficients vary by household, are then treated randomly, and follow a joint normal distribution. The marginal WTP for each energy source then varies by household and is also random.
Choice experiment and survey
Survey design
Energy WTPs for traditional and renewables are estimated by applying a hedonic price model. This requires survey data on energy prices by a combination of energy sources and characteristics of the energy combination, household, and household property. Our choice experiment survey resulted in a nationally representative sample of United States residents. Participants were asked to choose among three energy alternatives with varying costs. The first two alternatives consisted of a combination of two energy sources and the third allowed the households to keep their current energy source (primarily sourced from fossil fuels: oil, gas, and coal).
Our experiment includes three attributes, which together define various energy plans. The first attribute is an energy source, which is a blend of renewable energy (biomass, solar, and wind) with traditional energy. Households can choose to keep their traditional energy source. The second attribute is the percentage of a renewable energy source blended with the traditional fossil-fuel-generated energy. Levels 5, 15, and 20% are used for renewable energy sources. Only one renewable energy source is included in each option, with the remaining percentage being the traditional source. The third attribute is the percentage increase (5, 15, and 20%) in the cost of monthly residential energy. Because the contribution to utility from each energy source depends on its amount in the blend, the percentage variable and the source variable enter the utility only in interaction, i.e., Solar X percentage. As a result, Equation (2) can be written specifically as the following when the energy variables are continuously measured by percentage instead of dummies and their coefficients are the marginal utilities brought by a percent point increase in the energy sources. This is under the assumption that both price and energy share are linearly related to utilities.
$${V_{ij}} = \;{\beta _{iP}}{P_{ij}} + {\beta _{i1}}Sola{r_{ij}} + {\beta _{i2}}Win{d_{ij}} + {\beta _{i3}}Biofue{l_{ij}} + {\beta _{i0}},\,Traditiona{l_{ij}}$$
A full factorial design results in 33(33−1)/2 = 351choice sets. It is difficult to implement this magnitude of choice sets and obtain unbiased responses. We implemented a D-efficiency design yielding 60 choice sets divided into ten blocks. Figure 1 illustrates an example choice set. Each participant was asked to make one choice among A, B, and C options in each choice set and do this for six given choice sets. This represents the situation where households make a choice of energy produced with a particular percentage of a renewable source blended with fossil fuel at alternative price levels.
Figure 1. Example of a choice set.
Survey characteristics
We employed a major U.S. online survey company to conduct a national online survey using their panel in 2018. The company provided incentives for participants to opt in. Our U.S. representative sample yielded 2282 surveyed households with completed answers to all questions used in this study. Summary statistics are listed in Table 1. We compare our sample characteristics to the 2017 United States Census. Households are from all 50 states with proportions similar to state populations. Our sample is 69% female, with an average age of 46 years, and 75% white. These statistics are slightly different from the national average of 49% female, average age of 38 years, and 61% white. The annual household income of $51,982 is slightly lower than the national median household income of $63,688.
Table 1. Summary statistics of the whole sample (N = 2,282)
We asked households about their energy consumption characteristics: residential size; electricity bills during the spring, fall, winter, and summer months; major energy sources used to generate their electricity; number of central air-conditioning units; and the setting of their thermostat during winter months. The average residential size is approximately 1590 square feet and average monthly electricity bills range from $138 during the spring and fall months to $178 during the winter months. The survey indicates that 81% of households use electricity generated by traditional fossil sources: coal or oil (45%) and natural gas (36%). Only 1% of households reported using wind and biofuel and 3% reported using solar as their residential energy source.
Previous research indicates emotions play a significant role in supporting environmentally friendly public policies (Smith and Leiserowitz Reference Smith and Leiserowitz2014). Negative emotions concerning climate change may translate to prosocial behavior toward renewable energy. On a Likert scale, we asked households to rate the intensity of different emotions they felt while thinking about the issue of climate change. Emotional responses representing primary and emotions (Plutchik Reference Plutchik, Plutchik and Kellerman1980) included depression, guilt, worry, interest, disgust, sadness, anger, hope, helplessness, and fear. We conducted a principal component analysis (PCA) for an ordinal scale system to create an index (Emindex), which represented these ten emotions for each household. The index represented the range of negative responses toward global warming. A higher index magnitude represents an increased intensity of negative response toward climate change. The PCA correlation matrix of ten emotion scores is reported in Appendix Table A1. The pairwise correlations are all positive and large.
We also asked about households’ beliefs about the environmental impact of several electricity sources: wave, tidal, wind, geothermal, solar shingles, solar private panels, solar farms, coal, natural gas, hydro, biofuel, and battery storage owed by companies or by private homes. The survey indicates that 93% of households feel that the wind, private solar panels, solar shingles, and natural gas have a positive or neutral environmental impact. In contrast, only 85% feel that way about coal. The households’ perception of the environmental impact of coal is statistically different (p < 0.01) from their perceptions of the environmental impact of other energy sources. This implies among all sources, households considered coal as the least environmentally friendly.
Our survey also had questions to elicit households’ beliefs about the environmental impact of replacing fossil fuels with renewables. Results indicate 91% of households believe renewable energy sources will be the future; ordered by solar shingles, natural gas, solar panels, biomass, and biofuels. Strong agreement exists among households concerning wind and solar energy, although there is a small disapproval for different forms of solar energy. Specifically, households consider massive solar panels installed as solar farms inferior to privately installed solar panels or solar shingles. There is also a larger disagreement concerning bioenergy.
Results
Energy preferences
Table 2 lists the WTP summary statistics. The WTP standard deviations for all four energy sources are significantly difference from zero (p < 0.01). This supports our adoption of the heterogeneity hypothesis of different WTPs for renewables relative to conventional energy. Further, the magnitudes of the three renewable WTP standard deviations are similar, which indicates that the level of preference heterogeneity for each of the energy types is similar among the households.
Table 2. Marginal WTP for alternative energy of the whole sample (N = 2,282)
***, **, *denote significance at 1%, 5%, and 10% levels, respectively, with standard errors in parentheses.
All the mean WTPs are also significantly different from zero at the 5% level. The positive marginal WTP estimate for solar energy is 0.15. This implies if 100% of the energy for a household is from a solar energy source, then the mean household is willing to pay a 15% higher energy bill. Households perceive solar energy as a higher utility compared to conventional energy. The additional utility for solar energy is likely perceived from the public environmental benefits a clean source of energy brings and households’ prosocial satisfaction.
The estimated mean WTP for traditional energy is−7.1%. This implies in the absence of renewables if the only source of energy available to households is traditional energy, the households demand a discount of 7.1% off their energy bill. The estimates for the WTP for wind and biofuel are−11.5% and−65.1%, respectively. Households appear to dislike these two energy sources relative to traditional energy. Both wind and biofuel are controversial. In particular, the 65.1% discount for biofuel indicates a mean household demand to be paid for using biofuels. Their bias toward using food/feed for generating electricity and the questionable carbon footprint may explain their negative support for biofuel. The 11.25 discount of wind energy is smaller in magnitude with a lower significance level. These WTP estimates for renewable energy sources are consistent with previous studies (Borchers et al. Reference Borchers, Duke and Parsons2007; Aguilar and Cai Reference Aguilar and Cai2010).
WTP determinants
A WTP for each energy source is estimated for each household based on their response to six independent choice tasks. Regressing households’ WTPs on their attitudes, emotions, and socioeconomic characteristics, we can draw inferences on factors affecting their WTPs. In Table 3, results indicate younger households have higher relative WTP for the three renewable energy sources. This is consistent with research indicating younger households have heightened environmental concerns (Shi et al. Reference Shi, Visschers, Siegrist and Arvai2016). Similarly, households with higher income and education have enhanced WTP for solar and wind energy. These results are also consistent with general economic evidence that households with higher income often have higher WTP for products supporting environmental sustainability (Saphores et al. Reference Saphores, Nixon, Ogunseitan and Shapiro2007) and educated households are often more aware of the environmental issues and assign higher utility to these factors (Sundt and Rehdanz Reference Sundt and Rehdanz2015). Furthermore, people belonging to the White race have higher WTP for wind energy and households with a larger house have higher WTP for a biofuel source.
Table 3. Estimation results of factors affecting WTPs from the whole sample (N = 2,282)
***, **, *denote significance at 1%, 5%, 10% levels, respectively, with standard errors in parentheses.
The seasonal electricity expenditure is generally insignificant, except for the summer months, which is significant, at the 5% level, for wind energy WTP. Households with higher summer electricity bills might be more aware of the positive aspects of renewable energy sources and concerned about the environmental effects of traditional fossil fuels.
Compared to those who don’t know their current sources of energy, households knowing their current energy to be from fossil sources including coal, gas and oil show a high willingness to pay for all solar, wind, and biofuel, which is reasonable. In contrast, households with their current energy sourced from nuclear energy had lower WTP for wind power. Nuclear, itself is a type of energy that does not generate greenhouse gas nor emit any polluting emissions. Previous research indicates consumers are WTP for emission reduction via increased reliance on nuclear energy (Roe et al. Reference Roe, Teisl, Levy and Russell2001). Consumers with nuclear energy may not realize the benefit of adopting renewable energies for the sake of environmental protection.
The emotional index has a positive influence on households’ WTP for all three renewable energy sources. Households with heightened environmental and/or climate change concerns exhibit higher WTP for renewable energy sources. Our results are supported by previous research, where environmental concern is an important determinant of WTP (Kotchen and Moore Reference Kotchen and Moore2007; Ito et al. Reference Ito, Takeuchi, Tsuge and Kishimoto2010).
To understand the beliefs and perceptions of households concerning the different energy sources, we considered households’ perceptions of the positive environmental impact of energy sources. Households who believed that coal has a negative environmental impact had a higher WTP for all three renewable energy sources. We presented three alternative methods to harness solar energy: solar shingles, privately owned solar panels, and solar farms. Households believing private solar panels have a positive impact on the environment had higher WTPs for solar energy and wind energy sources, while the other two methods have no effect. Households who perceived positive impacts from biofuel energy on the environment had a higher WTP for biofuels.
We also examined households’ beliefs concerning the impact of an energy source for future energy. A belief in biofuel hurts WTP for wind and solar sources, while the belief biomass will help curb pollution has a positive impact on the WTP for all three energy sources. A positive belief in wind and solar shingles has a significant increase in WTP for solar and wind energy sources. WTP for energy sources also varies with households’ political status. Democrat households have higher WTP for both wind and solar energy sources. Those who live in the northeastern region of the United States have lower WTP for wind and solar energy.
The overall results indicate households perceive solar energy as the preferred energy source in terms of environmental protection followed by wind energy. Energy generated through biomass is quite controversial and has not received customer acceptance. The highest WTP occurs from younger, higher income, and better-educated households with coal and oil as a current energy source, greater concern for environmental and climate change risks, and a negative coal perception.
Robustness check
We have done a robustness check for the same analysis when excluding all households who don’t know their current source of energy. Using the subsample of 1157 observations, descriptive statistics, WTP estimates, and the regression analysis of factors affecting WTP are presented in Appendix Tables A2–A4.
The descriptive statistics of all variables for the subsample are very similar to those of the whole sample. The WTP estimates also have the same signs and significance except for that WTP for wind energy is negative but insignificant for this subsample compared with the significantly small negative size. The WTP for solar energy is slightly higher for those who know their current source, 0.267 instead of 0.15.
The determinants of the WTPs also remain quite similar. Notice for the sample without those who don’t know their energy source, we use those who know their energy source is biofuel as the default for those source dummy variables, so this affects the signs and significance of all dummy variables in this energy source category but no other variables. The significant variables remain highly similar with the same signs.Footnote 2
Government mechanisms for renewables
As governmental incentives are motivating a shift toward renewables, an accurate understanding of household prosocial behavior is required to develop optimal government mechanisms. From an electric utility perspective, understanding marginal increases in household electricity expenditures from marginal shifts toward renewables aids in efficient adjustments facing a dynamic electricity market. This also allows the government to better understand the demand for renewables required for developing mechanisms that mesh with household prosocial behavior.
As an application, considering the optimal government mechanism for solar energy, Liu et al. (Reference Liu, Colson and Wetzstein2018) estimated an optimal subsidy of 7.69 cents/kWh, which did not consider any possible household prosocial behavior. Mating this with our estimate of households’ WTP of 15% more for solar energy and a national average of 13.14 cents/kWh instead results in a 5.71 cents/kWh optimal subsidyFootnote 3 (EIA 2020). Assuming no motivational crowding (Frey and Jegen, Reference Frey and Jegen2001), this results in a 26% reduction in the subsidy.Footnote 4 Furthermore, for those who know their current energy sources, this subsidy saving rises to 44.4%。 In contrast, biomass and wind energy would not exhibit any similar prosocial reduction. When determining efficient policies and programs, policymakers should consider these varying levels of prosocial behavior across renewable energy sources.
Conclusion and policy implications
As we aim to migrate away from fossil fuels and promote energy sustainability, renewable energy must play an increasingly dominant role. However, renewable energy technologies are generally more expensive than fossil fuels, especially since they are still in their early stages of development. In order for renewables to be widely adopted, the government needs to provide incentives to encourage adoption. The magnitude of these incentives depends on households’ preferences for renewable energy and their associated demand. Any positive prosocial behavior toward renewables will affect the magnitude of any governmental mechanisms. Our results indicate that households have a significant prosocial behavior toward solar energy, which reduces the optimal solar subsidy. This prosocial behavior is not universal across renewable energies. For instance, biomass energy has an inverse social behavior (anti-social behavior), so any government incentives for biomass would have to overcome both a higher cost differential and anti-social behavior. If biomass is to become a relevant feature of our energy mix, educational programs on the nature of its sustainability that change household perceptions will be required. Wind energy appears to be relatively neutral regarding its associated social behavior.
To fully comprehend the relationship between prosocial behavior, governmental mechanisms, and renewable adoption, we need to examine how they interact with each other. We have explored the varying degrees of prosocial behavior across renewable energy source alternatives. However, we also need to take into account how governmental subsidies can potentially reduce (crowd out) prosocial behavior. We have pieced together a major component of this puzzle by empirically unraveling the different magnitudes of prosocial behavior across renewable alternatives. Missing is the magnitude of possible motivation crowding out where a government mechanism may reduce (crowd out) prosocial behavior. For example, households may be less willing to voluntarily pay more for a renewable energy source if they receive a government subsidy for using it, reducing their reputation for prosocial behavior.
Understanding these differences in prosocial behavior are not only interesting theoretically, but also for determining the future of our energy power generation. Similar to markets in general, household preferences play a significant role in shaping energy markets. The market effect on nuclear power following Three-Mile Island, Chernobyl, and Fukushima Daiichi incidences is a clear example of this. The future mix of renewables with fossil fuels will be determined by these household preferences. Failure to understand households’ preferences and prosocial behavior toward alternatives may lead to inefficient government mechanisms but possibly failure.
Data availability statement
Data is collected from human subjects and is protected under the Institutional Review Board (IRB) at Purdue University. It can made available on request to Purdue IRB.
Authors’ contribution
The first three authors formulate the study, develop the model, advise student coauthors in survey design and econometric analysis, and write the article. Jiang conducted the econometric estimation. Weiland designed and administered the survey primarily.
Funding statement
Authors acknowledge the National Institute of Food and Agriculture, United State Department of Agriculture (USDA), Power Plants rejuvenated with Biomass: An Economic, Social, and Environmentally Sustainable Transition to Clean Power grant funding (Program director and co-investigators).
Competing interests
The authors declare that they have no conflict of interest.
Ethical standard
Approved by the IRB of Purdue University.
Code availability
Can be available on request to the authors.
Consent to participate
A paragraph is included in the survey and has been approved by IRB of Purdue University.
Consent for publication
No survey participants’ individual information is in the publication. All coauthors agree to publish this manuscript.
Appendix A
Table A1. Correlation matrix of emotion variables
Table A2. Summary statistics for the subsample: knowing their energy sources (N = 1157)
Table A3. Marginal WTP for alternative energy of the subsample (N = 1,157)
***, **, *denote significance at 1%, 5%, and 10% levels, respectively, with standard errors in parentheses.
Table A4. Estimation results of factors affecting WTPs from the subsample (N = 1,157)
***, **, *denote significance at 1%, 5%, and 10% levels, respectively, with standard errors in parentheses.
Open access
