**1. Introduction**

In July 2010, Norway implemented the energy labeling system for houses and dwellings, and energy performance certification became fully mandatory. Since then, all houses and dwellings for sale are required to have an energy performance certificate (EPC). The motivation for using EPCs is to provide information to buyers and tenants about the energy performance of buildings. Reliable information on energy consumption is supposed to improve the functioning of real estate markets and create incentives to invest in energy e fficiency. The information provided to potential buyers by the EPC is intended to stimulate energy e fficiency investments because the consequent improved energy performance will potentially increase the sale prices and rents of buildings [1].

The EPC reflects the expected energy consumption of a building, which enables buyers to account for the expected current and future energy costs when assessing their willingness to pay for a residence. However, the expected energy costs are not only a function of energy consumption. Energy prices and the interest rate will also influence the energy costs, with the latter working through the discounting of future values into present values. Hence, energy performance will potentially influence the transaction prices of dwellings and houses, in combination with the influence of energy prices and the interest rate.

The empirical literature has drawn contrasting conclusions concerning the role of EPCs in energy conservation [2]. In the commercial segment, Eichholtz et al., found that US o ffice buildings with a "green rating" sold for prices about 16% higher than did those without such ratings [3]. In a study applying hedonic regression on residential dwellings in the Netherlands, Brounen and Kok found a price premium for houses labeled as more energy e fficient [4]. (A hedonic regression breaks down the house price into its constituent characteristics and obtains estimates of the contributory value of each characteristic.) Fuerst et al., used both hedonic and augmented repeat sales regressions, and found a significant EPC premium for dwellings sold in England [5]. In addition, a report to the European Commission concluded that EPCs have a significant e ffect on property prices and rents in selected European Union (EU) countries [1].

Other studies indicate that EPCs have a weak or negligible impact on transaction prices. Interestingly, Murphy investigated the case of the Netherlands, that is the same housing market as Brounen and Kok [4,6]. By applying an online questionnaire, she studied the role of the EPC in the transaction process of buildings. Contrary to Brounen and Kok, she concluded that few householders pay attention to the EPC and stated that the EPC would not have the planned impact, even if the system was fully implemented [4]. Similar surveys from the UK by Laine, and from Germany by Amecke came to the same conclusion, namely that EPCs only have a minor or negligible e ffect on price negotiations and investment decisions [7,8]. In a similar manner, Backhaus et al. performed in-depth interviews with homeowners in 10 European countries, as well as a large survey among homeowners in five European countries, and found that EPCs have a modest or negligible impact on homeowners' purchase decisions [9]. In a hedonic model for single family housing in Sweden, Wahlström found no price premium of EPCs, but rather a price premium for housing attributes that improve the energy efficiency [10].

Olaussen et al. carried out a statistical study resembling that of Brounen and Kok for the case of Norway [4,11]. However, they reached a similar conclusion to that of Murphy [6]. Performing a hedonic regression analysis based on housing transactions in Oslo, the capital of Norway, they concluded that there was no price premium caused by the energy label itself. Indeed, they suggested that the positive price premium of the EPCs found in the former studies was the result of the methodological design rather than evidence of the impact of EPCs.

Olaussen et al. took advantage of the fact that the EPC system was implemented in Norway by the governmen<sup>t</sup> "overnight" on July 1 2010. This meant they had a quasi-natural experimental design with pre- and post- EPC data [11]. For each dwelling that was sold before the implementation of the EPCs in 2010, they identified the energy label that the same dwelling was given when resold in 2014. Interestingly, when using the energy labels of dwellings resold in 2014 as a variable in a hedonic regression for dwellings sold before the introduction of the EPC system, they found the same positive relationship between energy labels and the transaction prices. This means that the positive price e ffect of the energy label was present even before it was implemented, which strongly suggests that the studies that found positive price e ffects from the energy label captured something other than the e ffects of the label itself.

However, Olaussen et al. and many other earlier studies on the impact of EPCs on transaction prices did not account for changes in the energy prices in their analyses [11]. Indeed, looking at the development of the energy price in Oslo, a trend is observed whereby the energy price increases, reaching a peak in 2010, and then decreases. As 2010 was the year in which the EPC was implemented in Norway, it is possible that the lower energy price in the post-label period actually neutralizes the potential price premium of the EPC implementation. To gain a more comprehensive understanding of the impact of EPCs on transaction prices, we include a time series for the energy price in the analysis.

Based on the energy price time series, as well as extracting information about the expected energy consumption of buildings from their assigned energy label, we calculate the expected annual energy cost of buildings. Moreover, following Olaussen et al., we utilize the fact that energy labels were implemented overnight on 1 July 2010 [11]. Instead of focusing on labels, we now focus directly on the energy consumption. Therefore, to each dwelling sold before the implementation of the EPCs, we assign the same expected annual energy use that was calculated for that dwelling when it was resold after 2010. Thus, using a hedonic regression, we can assess whether the post-label impact of the expected annual energy cost on transaction prices is stronger than the pre-label impact. Our results indicate that the impact of the expected energy costs on transaction prices is more moderate after the implementation of the EPCs.

The paper is organized as follows. Section 2 outlines the energy labeling system and some descriptive statistics. Section 3 describes the method, and the results of the empirical analysis are presented in Section 4. A discussion and concluding remarks are provided in Section 5.

### **2. The Energy Labeling System for Dwellings and Houses**

The energy performance of buildings directive (EPBD) is the EU's main legislative instrument to improve the energy performance of buildings [12]. Based on the EPBD, the EPC system was implemented gradually throughout di fferent member states from 2006. The final deadline for implementing an EPC system in the member states was 2009. A recasting of the EPBD [13] in 2010 strengthened the role of EPCs by demanding that when buildings "are o ffered for sale or for rent, the energy performance indicator of the energy performance certificate of the building or the building unit, as applicable, is stated in the advertisements in commercial media" [13] (p. 24), rather than at the time of signing a purchase agreemen<sup>t</sup> or rental contract [1].

In most EU member states, the energy performance ratings are expressed on a letter scale, for instance, from A to G, where A is very e fficient and G very ine fficient, and this is also the case in Norway. As improved energy performance of buildings are supposed to increase sales prices and rents, the EPC is intended to generate incentives among owners to invest in improving energy e fficiency [1]. Still, the implementation of EPCs has been slow in EU, and it has been argued that both the implementation and quality of certification schemes vary from country to country [1].

As mentioned in the introduction, the EPC system was fully implemented in Norway on 1 July 2010. The Ministry of Petroleum and Energy and the Ministry of Local Government and Regional Development were given overall responsibility for the introduction, with the Norwegian Water Resources and Energy Directorate (NVE) as the authority in charge of the certification and inspection schemes [14]. The EPC system was fully mandatory from the beginning, that is, from July 2010, and since then all transactions involving houses and dwellings in Norway must be accompanied by an EPC.

The EPC is a legal document that is required to be shown to the buyer. However, parts of the certificate, for example the Energy Label, can be used as a simplifying short version [14]. Among other details, the document contains data identifying the building, the agen<sup>t</sup> issuing the certificate, the energy label (which represents the calculated delivered energy need) on a scale from A to G, the heating grade (which represents to what extent heating of space and water can be done with renewable energy sources) represented by color grades, advice on how to save energy, and some general recommendations to the buyer [14]. (Since most Norwegian homes use electric heating, the Norwegian EPC system focuses on energy consumption.)

The operational liability of the EPC system in Norway is with ENOVA. (ENOVA is a state-owned company owned by the Ministry of petroleum and energy until July 2018 and then by the Ministry of climate and environment. ENOVA is responsible for funding cost e fficient changes in energy production and consumption.) For the owners of existing buildings, there is a self-assessment option in the certification scheme. Normally, these certificates are less detailed than those carried out by professionals, and the cost of the certification process for these buildings is typically at least NOK 1000 (NOK = Norwegian kroner). The hedonic regressions in the studies from the di fferent countries we have referred to in the introduction show surprisingly similar values regardless of whether the EPC scheme was based on self-assessments or assessments made by experts. This includes both the assessment of the quality of the dwelling with respect to energy e fficiency and the extra advertising costs associated with selling when energy label information is included. When it comes to new buildings, a qualified expert is required for certification and, hence, it is more expensive than for existing buildings. The quality assurance aspect of the Norwegian certification system is attended to by controls in the market, where wrong inputs may be considered a breach of contract. In such cases, a fine may be issued. The transaction process is supervised by the Norwegian Water Resources and Energy Directorate (NVE), which supervises whether EPCs are presented at sale, whether the EPCs reflect the actual energy efficiency standard of the building, and whether the experts meet the competence requirements [14].
