Chapter 3 Environmental Effectiveness: Data and Methodology
The introduction of NABERS [National Australian Built Environment Rating System] certification in Australia has made it possible to track measured site energy consumption over time in hundreds of existing office building assets. Prior to the commencement of voluntary energy certification in 1999, natural resource consumption was private information and, when voluntarily disclosed, was not fit for use in academic research because reporting and methods were ad hoc. By harmonising data collection methods, NABERS has made it possible to answer research questions associated with measured environmental performance outcomes in the Australian office building stock.
This chapter outlines the data and methods that will address two research questions associated with the introduction of NABERS certification outlined in Sections 1.1.3 and 1.1.4. The first of these questions is whether market differentiation via repetitive energy consumption auditing, the strategy employed by NABERS Energy to reduce greenhouse gas emissions, influences site energy use in existing office building assets. The method presented in this chapter calculates the effect per participant1 in the environmental effectiveness framework developed by Borck and Coglianese (2009), who argue that the product of participation rate and effect per participant, plus spillover effects accruing to non-participants represent the outcome of a market intervention (see Section 1.1.2). In this study, the intervention is the introduction of repetitive energy auditing, with the degree of intervention measured as number of certificates obtained. Models are then presented that will be used in Chapter 4 to estimate the change in energy consumption between two NABERS Energy certificates, one obtained in period s and a benchmark representing the first certificate obtained in period 1:
The second research question is whether energy performance outcomes per asset from a panel of voluntary adopters differ from a panel of mandatory adopters when comparing the two groups. As Chapter 1 described, this question is highly relevant to the formulation of public policy associated with greenhouse gas mitigation in existing commercial building assets. Chapter 2 discussed the policy context in Australia, where the BEED [Building Energy Efficiency Disclosure]
Act has recently mandated the advertisement of a NABERS Energy rating prior to sale and lease transactions in large office buildings. This chapter produces three definitions of mandatory adopter that are used in Chapter 4 to compare voluntary and mandatory adopter performance.
The first part of this chapter discusses the compilation of NABERS Energy data gathered to answer these two research questions. Section 3.2 describes this dataset statistically, while sections 3.3 and 3.4 develop empirical methods to evaluate the effect per asset for buildings in Australia that have obtained multiple NABERS certificates. A brief summary of the model specifications are provided in section 3.5. Chapter 4 presents and discusses the estimation of these models.
The primary data for this research are extracted directly from a comprehensive collection of ABGR [Australian Building Greenhouse Rating]2 and NABERS certificates issued over the past 14 years and compiled by the author. All publicly available ABGR certificates and NABERS certificates have been obtained from the internet (http://www.abgr.gov.au for early ABGR certificates and http://www.nabers.com.au for later NABERS Energy certificates). Certificates were obtained from these sites since the commencement of ABGR in 1999 up until April 2012. Additional NABERS Energy certificates issued between April 2012 and the end of October 2013 were obtained for assets complying with BEED Act disclosure regulations. These compliance documents, which will be referred to as “BEED Act certificates”, are published on the commercial building disclosure website (http://www.cbd.gov.au) and include a full NABERS Energy certificate. In summary, the full NABERS Energy dataset spans between August 1999 and October 2013.
To ensure a sufficient number of observations for robust statistical interpretation, only the popular NABERS Base Building Energy accounting scope is used. The alternate accounting scopes, Tenant Energy and Whole Building Energy, do not have sufficient panel data for statistical tests, so the term “energy dataset” in this thesis refers solely to a collection of Base Building ratings.
Figure 3.1 is an example NABERS Base Building Energy certificate. Two star ratings are provided; one excludes the effect of greenhouse gas reductions obtained through the purchase of Green Power offsets. To describe underlying performance data captured in an audit, each certificate includes the percentage of electricity purchased with Green Power offsets, a calculation of overall building energy use intensity in MJ/m2/year that is unaffected by Green Power, and two calculations of greenhouse gas emissions based on the Greenhouse Gas Protocol accounting framework (World Resources Institute and World Business Council for Sustainable Development, 2004). These raw data are converted into a “benchmark factor” that takes into account unreported background information: the number of hours per week the asset is in full operation and local climate indices. The specific calculation methodology of the benchmark factor has never been publicly disclosed. The benchmark factor is used to calculate the number of stars given to an asset, with a regional scale based on 2.5 stars as the median benchmark factor in each region of Australia. For a thorough methodology of raw data collection techniques and boundaries for NABERS performance audits, consult the guidelines provided to auditors (Department of Environment Climate Change and Water NSW, 2010).
After obtaining each certificate, it is necessary to organise and clean the data. Multiple certificates for the same asset with the same expiry date are removed to eliminate duplicates, with the chosen certificate having the highest NABERS identification number (a proxy for the issue date). A small number of certificates are missing data that clearly identifies the certified asset or were issued to represent performance across a portfolio of assets. As a result, 71 Base Building Energy certificates were removed from the dataset. In total, the cleaned dataset contains 3,661 unique Base Building Energy certificates. The certificates are then organised in issue sequence for each individual asset in the database based on ascending NABERS identification numbers. Table 3.1 shows there are 1,153 unique assets in the energy dataset, with 818 having been certified multiple times.
Energy Performance Data
In order to ensure valid comparisons, energy performance in this study is measured using the raw consumption data input into each performance rating. Looking at Figure 3.1, “Energy Intensity”, commonly abbreviated EUI for Energy Use Intensity, is intended to measure raw consumption. This metric is not altered by the decision to purchase Green Power offsets and has been consistently reported on every certificate; hence EUI is the logical choice for a comparative variable representing asset performance. Star ratings are associated with the benchmark factor, calibrated separately for each Australian city and thus unsuitable for comparison across regions. One must also assume that star ratings are unsuitable for comparison across time because it is unclear whether the undisclosed method for calculating a benchmark factor has changed over the 14-year life of the scheme.
Despite greenhouse gas mitigation being a key objective for investment in operational building energy efficiency, greenhouse gas emission figures from NABERS Energy certificates are not used. Early NABERS Energy certificates only report emissions that take Green Power offsets into account, leading to a number of “zero-emission” buildings. Second, early certificates exclusively use greenhouse gas accounting scopes one, two, and three (World Resources Institute and World Business Council for Sustainable Development, 2004), while later certificates switch between different accounting protocols. The presence of multiple accounting protocols leads to the discard of many valid certificates in an attempt to compare data only within the same accounting framework.3 Lastly, greenhouse gas accounting practice has been very dynamic over 14 years; even if accounting scopes were consistent, conversion factors between the raw data and greenhouse gas emissions are sure to have varied over time. Thus, it would be difficult to differentiate trends in greenhouse gas emissions between operational management and changes in accounting practice. Instead, using EUI, a metric associated with operational greenhouse gas management but unrelated to its accounting, is the logical choice.
A potential concern in the use of EUI to interpret greenhouse gas performance is that it limits the scope of greenhouse gas mitigation to operational energy efficiency investment. Owners wishing to improve NABERS Energy ratings have three options: invest in on-site operational energy efficiency, purchase Green Power offsets, or switch fuel sources to maintain energy consumption intensity while reducing source greenhouse gas emissions.
While this research is limited to measuring the first option directly, it will take into account the decision to purchase Green Power. All building owners electing to purchase over 1% of their electricity via the Green Power scheme in every NABERS re-certification are identified using a binary variable. Theory on the effects of offsets in environmental markets is mixed. While the intent of offsets is to optimise the costs of mitigating a public bad (Kotchen, 2009), Gans and Groves (2012) describe how offsets can substitute for mitigation, potentially increasing production of the public bad. While this thesis does not intend to make a significant contribution to greenhouse gas emission offset theory, it is possible to answer empirically whether regular consumers of Green Power use offsets as a complement to mitigation – meaning owners initially invest in on-site mitigation and turn to offsets once the net costs of on-site mitigation rise above the cost of offsets – or as a substitute to mitigation – i.e. owners purchase offsets in lieu of on-site mitigation. The BEED Act requires that buildings disclose NABERS Energy ratings without accounting for Green Power offsets, so one can expect offsets and on-site efficiency to be valued differently, with a premium on the latter.
The third strategy for owners interested in mitigating greenhouse gas emissions, switching fuel sources, appears to be rarely used in practice. The correlation from first to final certification of the ratio of greenhouse gases per unit of energy is above 0.9 in buildings that can be compared over time, which most likely reflects minor variance in accounting factors from year-to-year.
The disclosure of ownership details on each NABERS certificate enables the identification of “green owners”, which will be defined as owners that are explicitly differentiating their assets as green or sustainable in the property market. In the establishment of the Global Real Estate Sustainability Benchmark, Bauer et al. (2011) rated three Australian-based institutional owners – Stockland, GPT and the Commonwealth Property Office Fund – as three of the top five “Global Environmental Leaders” for publicly listed property companies. In addition, GPT and a fourth Australian-based direct property ownership company, Investa, were identified as the top two Global Environmental Leaders for private property holding companies. Australian property investors are also listed on other sustainability indices, such as the Dow Jones Sustainability Index Australia, but these indices are restricted to listed firms, while the Global Real Estate Sustainability Benchmark is most specific to property funds and includes both listed and unlisted firms.
While the Global Real Estate Sustainability Benchmark study provided a systematic approach to classifying owners, there is also anecdotal evidence that these four Australian owners have invested in resource efficiency since the introduction of NABERS. For example, the 2011 Annual Report from the Investa-managed ING Property Fund (Investa, 2011), shows that planned investment is inversely related to a NABERS Energy rating (see Figure P.1 in the Preface).
In this thesis, assets owned by the four Australian companies identified as Global Environmental Leaders by Bauer et al. (2011) – Stockland, GPT, Commonwealth Property Office Fund, and Investa – are identified using a binary variable indicating green ownership. One non-listed property investment company – Local Government Super, the firm featured in the introduction of this thesis – is also identified as a Green Owner due to their responsible investment strategy (Churchill et al., 2011). None of the foreign-based owners identified as Global Environmental Leaders from outside Australia were identified as holding NABERS-rated assets in Australia.
The process of assigning NABERS certificates to an individual asset makes it possible to generate variables based on the location of the asset. In particular, Australian four-digit postcodes convey two useful pieces of data. One is the state or territory each asset is located in. This is important because Australia has three distinct levels of government – federal, state and local – and certain states, including New South Wales, the Australian Capital Territory, and Victoria, were the earliest adopters of NABERS Energy. The New South Wales government continues to manage the NABERS certification process throughout Australia. Hence, the particular state location of an asset can proxy fixed state government effects that may influence the decision to invest in operational resource efficiency.
The second useful variable that can be generated from a postcode is whether or not an asset is located in a capital city central business district (CBD). Office markets in a CBD offer prospective tenants greater choice than smaller provincial or suburban centres. Greater competition between owners may lead to greater investment in resource efficiency in major cities as part of an asset positioning strategy. Postcodes are used to identify buildings located in each capital city CBD: 800 for Darwin, 2000 for Sydney, 2601 for Canberra, 3000 for Melbourne, 4000 for Brisbane, 5000 for Adelaide, 6000 for Perth and 7000 for Hobart.
Asset Size and Hours of Use
The locational variables described above may be limited in regard to clear interpretation. In particular, significance of the capital city variable – particularly when interacted with the state variable – could relate to a wide scope of fixed location effects, such as property market cycles, local laws, cultural variations, operating hours, or building size. Kok et al. (2012a) found that when attempting to explain the diffusion of energy efficient office assets in the United States at a particular point in time, lagged economic variables describing income, employment and property market conditions were highly correlated. Lacking enough NABERS observations in each market to control for differences in economic drivers of energy efficiency across 14 years of NABERS certification, this study leaves the state and capital city variables to proxy average economic variation. But it is possible to produce truncated datasets with two variables from external sources that may influence energy efficiency potential: asset size and hours of operation. For example, Scofield (2009) argues that smaller buildings account for a disproportionate share of energy-efficient buildings by asset count in the United States.
Three external sources are consulted to procure the net lettable area (NLA) of each multi-certified asset. First, BEED Act certificates contain data on the area of each individual tenancy in the building. If the scope of a BEED Act certificate is for an “entire building” (as opposed to “part building”), the sum of all tenancy areas is assumed to equal total asset NLA. For assets lacking an entire building BEED Act certificate, each NABERS certificate publishes the name of the owner, so websites of individual building owners were consulted for published information that included the NLA of their assets. Third, sales records published in biannual research reports from Colliers International in each major metropolitan area of Australia also report NLA of assets sold in that six month period. If NLA could not be determined from either of the two methods above, the NLA on these sales records was used.
When a choice was presented, only office NLA was calculated in mixed-use buildings. Note that to be eligible for NABERS Energy certification as an office building, the asset NLA must be 75% office space or greater, so the possible inclusion of small retail areas is not likely to influence the results. In total, asset NLA was obtained for 806 of the 818 multi-certified assets in the energy dataset.
Data on intensity of asset use is only available from BEED Act certificates. The measure of occupancy provided on the certificate is the value of “Rated Hours”, measured in hours per week, collected during the NABERS performance audit and used in the calculation of the NABERS benchmark factor. This figure represents the number of hours per week the building is “safe, lit, and comfortable for office work” (Department of Environment Climate Change and Water NSW, 2010). It is a useful measure of how intense the asset is used. One can expect assets with higher operating hours to consume more energy.
With a BEED Act certificate being the lone source of Rated Hours data, only 696 of 818 multi-certified assets have at least one valid observation. Because Rated Hours can change in response to operational management decisions, there were 20 assets with multiple BEED Act certificates reporting at least a 4 hour difference in Rated Hours over time. In these cases, as well as those with minor differences, an average value across all certificates with known Rated Hours was chosen to represent all certificates obtained for that asset.
As a result of the data gathering process described further in Chapter 5, additional exogenous variables are available for office buildings in central Sydney. In total, 119 of the 818 multi-certified assets in the database fall within the boundaries of the central Sydney office market as defined by the Property Council of Australia (Figure 3.2). Although this is a small sample, the use of a single market eliminates unobserved fixed effects variability caused by cross-market aggregation and, in this case, enables the research to test the effect of additional exogenous variables potentially omitted from the larger sample.
The additional data available for the Sydney subsample includes asset age and service quality rating. Asset age is obtained from the RP Data Cityscope database, and is calculated as the age of the asset, in years, at the end of 2011. The subjective quality of an asset is assessed by the Property Council of Australia (2006b) and explained in Section 2.3.3. Grades of “Premium”, “A”,”B”, “C”, and “D” are used to advertise these quality rating assessments and obtained for this research from a variety of agency reports. Owing to a small sample of C-grade assets, B- and C-grade assets are combined as “secondary” assets. There are no D-grade assets in the sample.
Table 3.2 provides a descriptive overview of the entire energy dataset with analysis by number of certificates obtained. To ensure sufficient sample sizes for each cohort, the number of multiple certificates is capped at 8. This means that 31 assets with more than 8 energy certificates are not analysed beyond their eighth certificate4. Note that the aggregate column on the far right only includes multi-certified assets; the column of assets with only one NABERS Energy certification is excluded from the totals.
The subsample of buildings that have an observed net lettable area (NLA) as described in section 3.1.4 is very similar to the entire dataset (only 12 observations are missing), so a description of the NLA data is provided in Table 3.2. However, there is a loss of 122 observations when the metric of Rated Hours is included. Table 3.3 presents the same descriptive statistics as Table 3.2, but for the subsample of 696 asset observations that include Rated Hours. Observations with one NABERS Energy certification are excluded because Rated Hours data was only collected for 15 of 332 assets and these single-certificate observations are not part of any investigation associated with Rated Hours data. The only notable difference between Tables 3.2 and 3.3 is a consistent increase in average asset size, which is to be expected because small assets are not required to obtain the BEED Act certificate that is the sole source of Rated Hours data.
Both tables of descriptive statistics suggest that participation in NABERS Energy is associated with a measurable improvement in asset energy efficiency on average. Mean energy consumption indicators decrease between a building’s initial certification and its final certification. Energy consumption variance also decreases from initial certification to final certification. The key variable of interest, change in consumption, shows a clear trend of increasing energy savings over time and a decrease in variance. Boxplots in Figure 3.3 demonstrate the reduction in energy consumption and variance, particularly the reduction of outliers, as the number of certifications increase. These boxplots also suggest that after five certifications, mean energy consumption begins to stabilise while variance continues to decrease. Methods to test the robustness of this trend will be presented later in this chapter.
Besides change in consumption, three variables are also associated with the number of certifications. As would be expected, the number of certificates earned is related to the year a building first sought an assessment; early adopters are the only buildings with the highest numbers of certifications. Second, the percentage of assets managed by a green owner increases as the number of re-certifications increase. Unsurprisingly, this means green owners are likely to be early adopters of NABERS. Finally, there is a positive association between asset NLA, the percentage of assets in a CBD, and the number of certifications. This confirms the expectation that the CBD variable is correlated with asset size and also suggests that early adopters are more likely to own large assets. Figure 3.4 displays the fraction of large assets (greater than 20,000 m2) commencing NABERS Energy certification in each quarter since its inception relative to small assets; the early adopters of NABERS clearly own a higher proportion of large assets relative to later adopters.
Table 3.4 presents the correlation matrix between these related variables as measured in the dataset of all assets with observed NLA (N=806). The strongest correlation is between year of entry into NABERS and number of certificates. As described above, there are cross-correlations between green ownership, building size and the number of certificates. Green owners begin NABERS certification early and are likely to own large properties. Hence the interpretation of the green owner variable needs caution because it could be measuring green strategy as intended, or it could represent unmeasured characteristics of large institutional property owners, such as greater access to capital or the involvement of professional property managers.
Table of Contents
Table of Contents
List of Tables
List of Figures
Glossary of Acronyms
Co-authorship Declaration Form
Chapter 1 Introduction & Literature Review
1.1 The Macro Context: Sustainability, Green Building and Energy
1.2 The Micro Context: Market Effects
1.3 Summary of Objectives and Hypotheses
Chapter 2 Green Building in Australia
2.1 Green Building Differentiation in Australia
2.2 Green Building in Australian Public Policy
2.3 Australian Office Markets
Chapter 3 Environmental Effectiveness: Data and Methodology
3.2 Descriptive Statistics
Chapter 4 Environmental Effectiveness: Results and Conclusions
4.3 Next Steps
Chapter 5 Sydney Lease Market Effects: Data and Methodology
5.2 Descriptive Statistics
5.3 Modelling Semi-Gross Face Rent
5.4 Modelling Effective Semi-Gross Rent
Chapter 6 Sydney Lease Market Effects: Results and Conclusions
6.1 Semi-Gross Face Rent Results
6.2 Effective Rent Models
6.3 Conclusions and Discussion
6.4 Next Steps
Chapter 7 Conclusions
7.1 Summary of Research Findings and Contributions
7.2 The Energy Efficiency Investment Enigma
7.3 Implications for Public Policy and Property Practice
7.4 Limitations of this Study
7.5 Future Research Directions
7.6 Concluding Summary
List of References
GET THE COMPLETE PROJECT