Heterogeneous Price Premiums in Sustainable Real Estate? An Investigation of the Relation between Value and Price Premiums

Abstract:
Focusing on the voluntary LEED and ENERGY STAR environmental certification schemes in the United States, we investigate whether price premiums exist across all building value categories or are localized to specific value segments. We find that the largest value building segment does not demonstrate any price premiums, while the smallest value categories do. The concentrated supply of eco-labeled offices in large, high-quality buildings likely contributes to this phenomenon. Results from hedonic and quantile regressions indicate that price premiums for eco-certified real estate assets may not be uniformly distributed across value segments and that price premiums found in the literature are concentrated in smaller and mid-tier value buildings. This may be due to the  comparatively lower market penetration of eco-certification schemes in these segments.

...

There is some evidence that the effects of eco-labeling may not be uniform across time and space. For instance, Fuerst and McAllister (2009) found that the positive effects on occupancy rates of the ENERGY STAR label were only significant for the two worst performing deciles....  a great deal of empirical research exists on whether environmental certification effects the prices of real estate assets (Chegut, Eichholtz, and Kok, 2014; Robinson and Reichert, 2015)....A price premium was established in many studies and summarized in Fuerst and McAllister (2011a). Broadly, the results for rental rates tend to be fairly consistent with estimated premiums of 3%–5%. Researchers in the United Kingdom and Australia also find similar positive correlations (Fuerst and McAllister, 2011c; Gabe and Rehm, 2014). Smith (2015) discusses the growing trends for green space in India.  However, the results for sale price effects have varied widely with different model specifications and samples... While this body of work is too large to review comprehensively, few have attempted to investigate the extent of cross-sectional and temporal variation in the price effects of eco-certification. Probably, the most consistent finding in the body of work examining spatial variations in the market penetration of eco-certified buildings is that there is a positive relation between affluence and the rate of adoption (Choi, 2010; Fuerst, Kontokosta, and McAllister 2011; Kok, McGraw, and Quigley, 2011). It is notable that there has been no investigation of supply and price effects. Drawing on the CoStar database, Miller, Spivey, and Florance (2008), Eichholtz, Kok, and Quigley (2010), Wiley, Benefield, and Johnson (2010), and Fuerst and McAllister (2011a) produced papers that examined the sale and rental price premiums of LEED and/or ENERGY STAR certified offices in the U.S. Pivo and Fisher (2011) found premiums using appraisal-based data. Reichardt, Fuerst, Rottke, and Zietz (2011) applied a difference-in-differences approach and panel regression to estimate variations in rental premiums over time. For the panel regression fixed effects model, average rental premiums of 2.5% and 2.9% are estimated over the sample period. Reichardt (2014) finds lower operating expenses for LEED buildings. Das and Wiley (2013) interacted  variables such as size, class, and age with LEED and ENERGY STAR dummies to assess whether there were any additive effects. They find mixed results with the strongest being an additive effect on the price premium associated with size. Dippold, Mutl, and Zietz (2014) find differences in local demographics may effect green certified stock, while Freybote, Sun, and Yang (2015) find  neighborhood certifications may impact property types in that neighborhood. Some effect of green certification has been noted on mortgages in the residential and commercial space (Pivo, 2014; Sanderford, Overstreet, Beling, and Rajaratnam, 2015)....

Taipei 101, the tallest and largest green building of LEED Platinum certification in the world
https://en.wikipedia.org/wiki/Green_building

Eco-buildings are differentiated by ENERGY STAR, LEED, or dual as described earlier. Previous studies indicate different eco-premiums for each of the three labeling categories and they are thus treated distinctly. For Model 1, we find that ENERGY STAR, LEED, and dual certified offices sold at large premia of approximately 24%, 32%, and 32% respectively compared to non-certified offices. However, when these are weighted by price, these estimated premia decrease dramatically falling to 6%, 13%, and 9% respectively. The value-weighted results for the three segments are, therefore, consistent with the interpretation that premiums for eco-certified offices tend to be associated with smaller, lower value, buildings. This finding motivates further investigation into building price segmentation to determine what strata of building value drives price premia.

Models 3 and 4 are based on properties that sold for less than $26 million, which accounted for the vast majority of the observations from this data set of 10,000 SF office buildings and larger (22,993 out of 25,417 observations). It is only for this smallest value segment that statistically significant premia are identified for eco-certified offices. Models 5–8 show no price premia for any of the eco- buildings. This finding of small building premia is only consistent with the initial value weighting. When each observation is no longer treated equally, but weighted by their respective dollar value, the premia reduces. 

...

The rental results of a regression on rent PSF in Exhibit 5 show similar patterns to the sales, albeit not identical. Again the control variables are as expected in the whole sample regression with size generally increasing revenue, triple net lease structure subtracting rent, and percentage leased  increasing PSF rent as well. The ENERGY STAR buildings show premiums for the whole sample (Models 1 and 2) and in the smallest gross revenue category (Models 3 and 4), but not for the middle and larger value buildings (Models 5–8). This is a consistent pattern as with the sales sample. Curiously, revenue weighting increases the premia in the whole sample regression for ENERGY STAR, but the difference is economically small.

LEED only buildings, as in the sales models, do not show premiums in the largest value category. This finding of no rent premia amongst the highest revenue (value) buildings is further evidence of a de facto requirement for green features in the large, high value asset segment. LEED buildings do show rental premiums in the middle section as well as the smallest. The dual buildings are  disproportionately found in the largest value section and do show a rental premium of 6.3%.

Interestingly, when revenue weighted in the largest section, the premium increases dramatically. This could be because of a small number of super-A, trophy buildings demonstrating rental premiums.  Their large square footage even further impacts on weighting. With the exception of the dual  category, likely driven by some super-A dual buildings as suggested by the weighted regression, the general pattern of smaller revenue buildings showing premiums and larger buildings not showing them holds in the rental data.

The full article is available free of charge at http://www.josre.org/wp-content/uploads/2015/11/7.01-Heterogeneous.pdf.

by Spenser Robinson and Pat McAllister

Journal of Sustainable Real Estate JOSRE www.JOSRE.org

Volume 7, Number 1; November 12, 2015