Energy Efficiency and Energy Governance: A Stochastic Frontier Analysis Approach.

AuthorBarrera-Santana, J.

    Improving Energy Efficiency (EE) is a primary objective in the energy policy in OECD countries, as it is one of the most cost-effective ways to reduce greenhouse gases and energy dependence, and to guarantee the security of supply (IEA 2019). However, the effectiveness of these policies is conditioned by the existence of regulatory failures, an improper coordination between the government and the private sector or the lack of specific laws that correctly incentivize improvements in EE (Gillingham, Newell, & Palmer 2009). For this reason, the International Energy Agency (IEA), in its report from 2010, highlights that the success of energy policies aimed at improving EE is closely related to the performance of energy governance (IEA 2010). Furthermore, the report emphasizes the importance of analyzing and measuring energy governance (more specifically, EE governance) separately from the general governance of a country, highlighting up to three main areas: the regulatory framework, institutional agreements and coordination mechanisms of EE policies. (1)

    The impact of general governance on the economy has been widely studied in the literature (Acemoglu, Johnson, & Robinson 2001, among many others), whereas the effect of energy governance on the energy sector has been analyzed to a much lesser extent (Pereira & Pereira Da Silva 2017). This fact can be partially explained because general governance is properly measured by well-known indexes (e.g., World Governance Index--WGI--or Polity IV), while there is rather scarce quantitative information on energy governance. (2) In this sense, following IEA (2010), in Barrera-Santana, Marrero & Ramos-Real (2020) we built a composite index of EE Governance (the EEGI) for 32 OECD countries and representative for the period comprised between 2000 and 2015, using information about strategic plans, laws and decrees, or financing mechanisms, among others. The objective of this work is to use the EEGI to quantify, for the first time in the literature, the impact of energy governance on EE in OECD countries for the aforementioned period. (3) Furthermore, we will also compare our results with the most commonly used efficiency measure in the literature: Energy Intensity (EI).

    Our work draws on the theoretical framework developed by Filippini & Hunt (2011), who adapt the parametric Stochastic Frontier Analysis (SFA) proposed by Aigner, Lovell, & Schmidt (1977) to energy demand. The result is a frontier that represents the optimal level of derived demand for energy, measured through aggregate primary energy consumption. The location of each country with respect to such a frontier provides a measure of EE.

    The SFA has acquired remarkable relevance in the last decade in the field of energy (Tovar, Ramos-Real, & de Almeida, 2015; Filippini & Hunt, 2016). (4) On the one hand, it is an alternative strategy to the use of EI as a measure of EE, since the SFA overcomes the criticism made of EI (see Adom, Amakye, Kwabena Abrokwa, & Quaidoo, 2018). (5) On the other hand, this methodology imposes a functional form on the demand function, which provides three fundamental advantages very convenient for our application: (i) the efficiency estimates are more precise with respect to other alternative procedures (Filippini, Hunt, & Zoric 2014); (ii) the SFA reduces bias due to unobserved heterogeneity (Greene, 2015); (iii) external factors (e.g., the EEGI) can be included in the parametric function to explain EE (Filippini & Hunt 2015).

    We introduce the EEGI as an explanatory factor of the EE level in the three SFA specifications most widely used in the literature: Battese & Coelli (1995) (BC95), Greene's fixed effects (TFE) model (Greene, 2005a) and Greene's random effects (TRE) model (Greene, 2005b). However, since these methodologies show relevant concerns that should be addressed, we introduce three alternative approaches: the Marginal Maximum Likelihood Estimator (MMLE), proposed by Chen, Schmidt, and Wang (2014); the Marginal Maximum Simulated Likelihood Estimator (MMSLE) and the Pairwise Difference Estimator (PDE), both proposed by (Belotti & Ilardi 2018). (6) As far as we are aware, these last two approaches have not been applied before to estimate an EE frontier.

    Our results show a highly significant and positive effect of energy governance on EE. Increasing the EEGI score by 10%--an amount that can be interpreted as an acceptable goal in the medium or long term in OECD countries--could raise the EE level, on average, around 9.20% according to the PDE model and up to 27% in TRE. (7) In the most efficient countries, energy governance is characterized by the presence of both measurable (i.e., quantified) targets and extensive evaluation mechanisms. To a lesser extent, these countries also show a broad number of public-private co-operations, a well-developed regulatory framework and energy providers highly involved in energy policy. (8)

    This work contributes to the literature on energy economics in several aspects. First, we provide an unprecedented and thorough assessment of the role of energy governance (and different specific energy policies) in the EE level for a large sample of OECD countries. Secondly, the work also contributes to the definition and estimation of EE through the SFA, since the use of our EEGI improves the accuracy of the econometric models by reducing the bias due to the heterogeneity of unobserved variables. Moreover, we contribute in the use of novel SFA approaches (Chen et al. 2014; Belotti & Ilardi 2018), which overcome some of the problems raised by traditional methods, and also in the endogeneity assessment between relevant variables of the analysis, which is a relatively new concern in the SFA literature (Karakaplan & Kutlu 2017; Karakaplan 2017).

    The remainder of this paper proceeds as follows. In Section 2 we summarize the construction and the results of the EEGI for a sample of 29 OECD countries. In Section 3 we present our SFA approach. In Section 4 we estimate the effect of the EEGI on EE drawing on different SFA models. In Section 5, we assess different indicators included in the construction of the EEGI, to analyze the most effective energy policies for improving the countries' efficiency. Finally, we provide a set of final remarks.


    This section describes the construction of the EE Governance Index (EEGI) that we use later to measure the impact of energy governance on EE. The EEGI is based on our previous work (Barrera-Santana et al. 2020), for a reduced sample of 29 OECD countries in the same period 20002015. (9) We provide further details in Appendix A.

    2.1. Theoretical basis

    The EEGI draws upon the theoretical framework provided by IEA (2010), in which EE governance is structured in three main areas. The first area is Enabling Frameworks, which includes the legal basis for the implementation of EE policies, including some financial aspects, which makes it a fundamental area that influences and conditions the other two. The second area, Institutional Arrangements, refers to the development of practical instruments that reflect the degree of intervention and implication of the different agents that participate in the sector. Finally, the third area, Co-ordination Mechanisms, considers the mechanisms needed to coordinate the different agents of the energy sector, as well as to monitor and evaluate the results of the measures applied to promote EE. In turn, each of these areas is composed of several indicators (see Figure 1).

    Due to external reasons--being the lack of available information the most relevant--the EEGI is able to asses 8 indicators out of a total of 12. Those of Resourcing Requirements, International assistance and Governmental Co-ordination have been excluded. In Appendix A, we provide a more detailed explanation for the exclusion of these indicators, and we also conduct different robustness assessments in order to demonstrate that such an exclusion does not significantly affect our main results.

    2.2. Data collection, scoring procedure and index construction

    To construct the EEGI, we use information from the IEA's Energy Efficiency Database (2016). These data contain national measures (e.g., policies, laws) on EE, in force in the period 2000-2015. Each record contains mostly qualitative information that may be related to one or more of the indicators shown in Figure 1. (10) A rigorous analysis was carried out to match the information with each indicator and area. Taking the original information as a reference, we created our own database with almost 1,700 records.

    The qualitative nature of the data requires the establishment of a consistent scoring scale and objective evaluation criteria. Based on the public investment efficiency index proposed by Dab-la-Norris, Brumby, Kyobe, Mills, & Papageorgiou (2012), we used a rating scale between 0 and 4, with the highest score corresponding to a better quality of EE governance. Depending on the indicator, we adapt the scoring criteria followed by Dabla-Norris et al. (2012) according to the amount of information available. For the sake of simplicity, the concrete explanation of how we score some indicators is described in Appendix A.2. For additional details, see also Appendix I in Barrera-San-tana et al. (2020).

    Once all the indicators have been rated, they are added considering the structure in Figure 1. Following the recommendations of the OECD and JRC (2008); Knack, Halsey Rogers, & Eubank (2011) and Dabla-Norris et al. (2012) the indicators for each area are averaged giving rise to three energy governance sub-indices, one representative for each area. Subsequently, these sub-indices are averaged to obtain the EEGI, which consists of a single value, from 0 to 4, representative for each country and for the period 2000-2015.

    2.3. Energy Efficiency Governance results

    Table 1 collects the detail of the EEGI scores, as well as the scores...

To continue reading

Request your trial

VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT