Household internal and external electricity contract switching in EU countries

ABSTRACT

Using a representative sample of more than 13,000 households from eight countries in the European Union (EU), this article empirically studies the factors related to household electricity contract switching by distinguishing between internal switchers (households that switched contracts but stayed with the same supplier) from external switchers (households that switched to a new supplier). The econometric analysis includes individual preferences, household structural factors and socio-demographic characteristics, as well as electricity market characteristics. The study explicitly explores the role of risk and time preferences on switching behaviours, with risk and time preferences elicited through incentivized experiments as well as self-assessment scales. The main results suggest that internal and external switching are not related to the same factors, that risk and time preferences affect switching behaviours, and that renters are less likely to switch than homeowners; further, electricity market characteristics are found to affect household electricity contract switching.

KEYWORDS:

• Electricity supplier switching
• inertia
• liberalization
• risk and time preferences

JEL CLASSIFICATION:

• D10
• D90
• Q40

Acknowledgements

Part of this work was funded by the EU Horizon 2020 Framework Programme under the project BRISKEE-Behavioral Response to Investment Risks in Energy Efficiency (project number 649875; http://www.briskee-cheetah.eu/). We are grateful for the comments we received from an anonymous reviewer. We also thank Katharina Rehfeld for her assistance in reviewing the literature and Stacy Malek for her thorough proofreading of the article.

Disclosure statement

No potential conflict of interest was reported by the authors.

Supplementary material

Supplemental data for this article can be accessed here.

Notes

¹ Thus, high shares of the network component and of taxes and other levies (e.g. to finance support for renewable energy sources) of the total retail price mean that retail competition only refers to a relatively small share of the total electricity bill.

² See also A1 in the Appendix for key characteristics of the electricity retail markets in the countries included in the empirical analysis in this article.

³ For the UK, Wilson and Waddams Price (2010) found that some consumers fail to select the most beneficial electricity contract. At least one out of five consumers chose a contract that made them worse off than before switching.

⁴ Shin and Managi (2017) provide a comprehensive review of the literature on the significance and the determinants of electricity-provider switching.

⁵ We note that using a nested model to estimate household electricity contract choice is not feasible, since information on the characteristics of the old and new contracts is not available. In such a model, the household is assumed to first decide whether to switch the electricity contract, and then whether to switch internally or externally.

⁶ Since the sampling method excluded individuals older than 65 years and households without internet access, the reported switching rates may overstate the population switching rates since the excluded groups are less likely to have switched electricity contracts.

⁷ The subsequent analyses implicitly assume that (the arguably few) respondents who are both internal and external switchers nonsystematically chose response category (2) or (3).

⁸ δ = 1 / 0 < δ < 1 means that the participant is not discounting future outcomes / discounting future outcomes.

⁹ To avoid order bias, we randomized the order of the decisions presented to participants. Thus, participants had equal chances of seeing AB and BA.

¹⁰ The monetary amounts displayed to participants were adjusted across countries with different currencies to keep the relative value similar in terms of purchasing power. The following rates were applied: Poland: 1€ = 3 PLN; Romania: 1€ = 3 RON; Sweden: 1€ = 10 SEK; UK: 1€ = 1£. In all Euro-zone countries, the monetary amounts shown to participants were identical. In addition, for about 10% (7%) of the total sample, all values shown in the MPLEs were multiplied by 10 (divided by 10), relative to the baseline treatment (medium stakes).

¹¹ Information on household actual electricity consumption or electricity costs was not available.

¹² Country-level data was obtained from the EUROSTAT database on electricity market indicators; for the number of households, we used the variable coded lfst_hhhnhtych; the main electricity retailors are those with a market share of at least 5% in a given country; the figure on retail_concentration for Germany was taken from the Bundesnetzagentur database (Bundesnetzagentur 2017) because it was not available in the EUROSTAT database.

¹³ The multinomial probit model was estimated using the mprobit command implemented in Stata 14, which assumes zero correlation of the stochastic components of the choice alternatives. We show the marginal effects on the probability of choosing a particular contract rather than the coefficients of the model output since the latter refer to the latent utility and are therefore hard to interpret. In addition, these coefficients depend on the type of contract chosen as the base outcome in the multinomial model. For dummy variables and z-score transformed variables (i.e. nonmarginal changes), Table 5 reports the discrete changes in probabilities.

¹⁴ All results that are not shown to save space are available from the authors.

¹⁵ This conclusion implicitly presumes that trust towards providers is related to risk perceptions (e.g. about security of supply).

Additional information

Funding

Part of this work was funded by the EU Horizon 2020 Framework Programme under the project BRISKEE-Behavioral Response to Investment Risks in Energy Efficiency [project number 649875; http://www.briskee-cheetah.eu/].