Elsevier

Energy Policy

Volume 35, Issue 6, June 2007, Pages 3327-3334
Energy Policy

Does willingness to pay for green energy differ by source?

https://doi.org/10.1016/j.enpol.2006.12.009Get rights and content

Abstract

We present the findings of a choice experiment designed to estimate consumer preferences and willingness-to-pay (WTP) for voluntary participation in green energy electricity programs. Our model estimates WTP for a generic “green energy” source and compares it to WTP for green energy from specific sources, including wind, solar, farm methane, and biomass. Our results show that there exists a positive WTP for green energy electricity. Further, individuals have a preference for solar over a generic green and wind. Biomass and farm methane are found to be the least preferred sources.

Introduction

The United States generates its electricity using a wide variety of fuel sources, primarily oil, natural gas, nuclear, and coal. Of increasing concern are the adverse environmental impacts of these energy sources. Due to pollution externalities these traditional energy sources tend to have market prices below their true social cost. Green energy sources, in contrast, have higher market prices than traditional sources but, likely, lower social costs. In this study, green energy includes solar, wind turbines, biomass, and farm methane. There appears to be real and growing interest among many consumers for environmentally friendly energy production. These preferences may make it privately optimal for some consumers to pay a voluntary premium.

Existing research reports positive willingness to pay (WTP) for green energy electricity premia. These studies elicit WTP for various aspects of green energy, where “green energy” is a generic product (Byrnes et al., 1999; Ethier et al., 2000; Gossling et al., 2005; Zarnikau, 2003) or focus on the environmental attributes associated with green energy (Bergmann et al., 2006). For instance, Zarnikau (2003) found 50% of respondents in Texas were WTP at least one dollar per month to support renewable and energy efficiency investments. In contrast, Bergmann et al. (2006) used preferences for environmental attributes to infer preferences for green energy sources. Roe et al. (2001) estimated WTP for inputs and outputs associated with green energy in a multiattribute setting, including non-price attributes such as changes in air emissions, contract terms, and fuel mix (combination of traditional and green sources). Although fuel mix was part of the design in Roe et al. (2001), preferences and WTP for individual green sources were not estimated.

Despite the evidence of WTP for green energy, existing green power programs have shown a median participation rate of only 1.0% (Bird and Brown, 2005). Limited participation may arise from a failure in marketing research; previous studies may have suffered from a stated-preference bias or other error and overestimated WTP for these premia. Alternately, an education or communication failure—high information costs—may exist between producers and consumers so that true demand for green energy does not materialize. This paper is motivated by a third possibility. Consumers may have evaluated the products available, in terms of price and source, and decided not to purchase because the product offered is a generic green energy good or is perceived to be an inferior type of green energy when consumers want to reveal demand for specific or superior green energy source.

This paper builds upon Roe et al. (2001) and estimates preferences for specific green sources and offers an additional empirical test about whether preference varies between specific and generic green energy sources. A contingent choice experimental design is used to examine a specific set of green energy attributes to better estimate consumer WTP. Data are collected from a sample of New Castle County, Delaware, residents to test the hypothesis that consumers distinguish between a single, generic “green energy” source—as modeled in existing literature—and four specific green fuel sources. In addition, the results estimate the marginal WTP for four green energy sources. The results show that there exists a positive WTP for green energy electricity. Further, the specific green energy source affects WTP. In fact, individuals do exhibit preferences for solar versus a generic green or wind source. Biomass and farm methane are found to be the least preferred sources.

Changes in consumer welfare for the addition of green energy are estimated for two potential green energy program administration scenarios. First, changes in welfare are estimated for a voluntary program then a mandatory green energy program. The welfare estimation under these two scenarios results in different values. These results show the substantive significance of recognizing the variation in preferences for green energy sources.

The broader applicability of the results may be limited by several factors. Green energy preferences best reflect the geographical location where the data were collected. The sample was of small to moderate size. In addition, the choice was particularly salient at the time of enumeration given a recent large electric rate increase in the county. Nevertheless, the results are clear, and the findings may help to guide research and policy regarding green power programs in other regions.

The second section of this paper describes our model of consumer choice. The third section explains the experimental design, survey methods, sample, instrument, and descriptive statistics. The fourth section presents the econometric results and WTP estimations. A final section offers conclusions.

Section snippets

Conceptual model

This paper reports results from a contingent choice experiment administered through an intercept survey designed to examine preferences for green energy programs. These programs would deliver electricity to respondents via alternative green energy sources (wind, solar, biomass, farm methane) at some added cost to their monthly electricity bill. Respondents were asked to make choices between two green energy alternatives with varying attribute levels or to choose the status quo at no increase to

Results

LIMDEP software was used to estimate the NL model. Descriptive statistics are in Table 2 and NL results are shown in Table 3. Overall, the estimated NL model performs well, and the null is rejected using a chi-squared test (p<0.0001) and the model has a pseudo R2=0.19. The estimated inclusive value coefficient can be used to determine the appropriateness of the NL model. The test that the inclusive value is one can be rejected at the 1% level, which suggests that the single nest or MNL model

Conclusions

This paper presents the results of a choice experiment designed to elicit WTP for green electricity and to determine respondents’ relative preferences for specified green sources. The results suggest that respondents do exhibit positive WTP for green energy in general and for each of the specified green sources. However, some specific green sources are more preferred than others.

The results also show that there is a difference between voluntary or opt-in programs and non-voluntary programs.

References (20)

There are more references available in the full text version of this article.

Cited by (326)

View all citing articles on Scopus
View full text