1# RECORD e06a4d3e776b1e6666c940f56997612ef11873e42# TITLE: The effect of BEAM Plus certification on property price in Hong Kong3# YEAR: 2017.0 VENUE: nan DOI: 10.3846/1648715x.2017.14092904# AUTHORS: E. Hui; C. Tse; Ka Yu5# PAGES: 17 CANDIDATE TABLE PAGES (1-based): [9, 13]6# ABSTRACT: In response to the public’s increasing awareness of sustainable development, the construction industry has introduced “green” buildings which emphasize better environmental performance. However, as a building’s environmental performance is difficult to discern for laymen, different green building certifications have been established for evaluations in this regard. This study evaluates whether there exists a price premium for living space in buildings certified with BEAM Plus. The findings show that the prices of flats in BEAM Plus-certified buildings are 4.4% higher than those in nonregistered buildings, and that housing units in buildings with an “unclassified” rating are transacted at a discount of 5.9%. Nevertheless, if homebuyers mistake an “unclassified” building for a non-registered building, due to the non-disclosure of the “unclassified” result by developers, the price premium of BEAM Plus certification becomes higher (6.2%). Regardless, the price premium is much lower than those on the office sector. The reasons behind such differences can be attributed to the disparities in the tangible and intangible benefits associated with green living space and green office space. Policy implications with reference to Hong Kong’s GFA concession policy are then discussed.789===== PAGE 1 (score 9, numeric-paren 1, table-words 0, var-words 8) =====1011Copyright © 2017 Vilnius Gediminas Technical University (VGTU) Press12http://www.tandfonline.com/TSPM13INTERNATIONAL JOURNAL OF STRATEGIC PROPERTY MANAGEMENT14ISSN 1648-715X / eISSN 1648-9179152017 Volume 21(4): 384–40016doi:10.3846/1648715X.2017.140929017* Corresponding author. E-mail: bscmhui@polyu.edu.hk18THE EFFECT OF BEAM PLUS CERTIFICATION ON PROPERTY PRICE IN 19HONG KONG20Eddie C. M. HUI a,*, Cheuk-kin TSE a, Ka-hung YU a21a Department of Building and Real Estate, The Hong Kong Polytechnic University, Hong Kong, China22Received 23 June 2015; accepted 2 March 201723ABSTRACT. In response to the public’s increasing awareness of sustainable development, the con-24struction industry has introduced “green” buildings which emphasize better environmental perfor-25mance. However, as a building’s environmental performance is difficult to discern for laymen, different 26green building certifications have been established for evaluations in this regard. This study evaluates 27whether there exists a price premium for living space in buildings certified with BEAM Plus. The find-28ings show that the prices of flats in BEAM Plus-certified buildings are 4.4% higher than those in non-29registered buildings, and that housing units in buildings with an “unclassified” rating are transacted at 30a discount of 5.9%. Nevertheless, if homebuyers mistake an “unclassified” building for a non-registered 31building, due to the non-disclosure of the “unclassified” result by developers, the price premium of 32BEAM Plus certification becomes higher (6.2%). Regardless, the price premium is much lower than 33those on the office sector. The reasons behind such differences can be attributed to the disparities in 34the tangible and intangible benefits associated with green living space and green office space. Policy 35implications with reference to Hong Kong’s GFA concession policy are then discussed.36KEYWORDS: Hong Kong; Green building certification; BEAM Plus; Residential property price37SUPPLEMENTARY MATERIAL associated with this article can be found, in the online version, at 38https://doi.org/10.3846/1648715X.2017.1409290391. INTRODUCTION40As the notion of sustainable development has be-41come increasingly popular since the early 1990s 42(Gibbs et al. 1996), more customers have placed a 43higher degree of importance on environmental pro-44tection than on economic growth (Mainieri et al. 451997). As a result, the demand for goods and ser-46vices with lower environmental costs has soared. 47To put it differently, there exists a willing-to-pay 48(WTP) premium for these products over regular 49products. In response, companies in various busi-50ness sectors have begun to take environmental 51considerations into account in their business op-52erations (Berry, Rondinelli 1998), as well as pro-53duction and marketing plans (see Hawken 1993; 54Stigson 1998; Gonzalez-Benito, J., Gonzalez-Beni-55to, O. 2005). This results in an expanding market 56for eco-friendly products (and services) on a global 57scale (Fuerst, McAllister 2011).58Such emphasis on the preservation of the envi-59ronment has generated an even more profound im-60pact to the practices of the construction industry, 61given the environmental impact incurred in the 62daily operation of buildings. Hong Kong’s electric-63ity consumption has been gradually rising in the 64last decade (Fig. 1). In 2014, the vast majority of 65electricity was used for the operation of either com-66mercial buildings or residential buildings (Fig. 2), 67in particular on space conditioning (Fig. 3). The 68soaring electricity consumption has serious impli-69cations on the environment, in that electricity gen-70eration alone contributed to over 68% of all green-71house gas emissions in Hong Kong in 2013, accord-72ing to the Environmental Protection Department. 73Based upon these statistics, it is reasonable to say 74that, should buildings be more energy efficient, 75greenhouse gas emissions in Hong Kong could be 76significantly reduced. In view of this development, 77the concept of “green buildings”, which emphasizes 78eco-friendly building design and higher efficiency 79in the use of energy, water, and other resources, 80has been preached as the blueprint of sustainable 81development, and used as one of the major market-82ing tools for development projects in Hong Kong.838485===== PAGE 9 (score 14, numeric-paren 0, table-words 2, var-words 8) =====868739288E. C. M. Hui et al.89Of the 646 housing units in the data sam-90ple used for this study, 54.3% of them are pre-91sale flats and the remaining 45.7% are occupied 92flats. As seen in Table 5, for the full sample, the 93transaction price of the housing units in the sam-94ple ranges from 2.2 million HKD to 9.13 million 95HKD, with an average of approximately 4.65 mil-96lion HKD. The majority of them were transacted 97at more than 4 million HKD (60.1%). The aver-98age floor area of these flats is 359.6 sq. ft, with 99the smallest housing unit at 193 sq. ft. and the 100largest flat at 717 sq. ft. More than 60% of the 101Table 5. Descriptive statistics of residential units in the sample102Full sample103Pre-sale flats104Occupied flats105Transaction price (in HKD)1063 million or less1078.21080.910916.91103.01–4 million11131.711224.511340.2114More than 4 million11560.111674.611742.9118Maximum (in million HKD)1199.131209.131218.70122Minimum (in million HKD)1232.201242.901252.20126Mean (in million HKD)1274.651284.961294.28130Standard deviation (in million HKD)1311.421321.371331.79134Floor area (in square feet)135200 or less1363.91377.11380139201–40014063.814168.114258.8143401–60014429.214524.814634.5147More than 6011483.114901506.8151Maximum152717153577154717155Minimum156193157193158241159Mean160359.61161331.32162393.16163Standard deviation164114.7316596.08166125.66167Floor level1681st–10th16930.117029.317131.117211th–20th17335.217438.817531.117621st–30th17728.717830.517926.718031st–40th1814.21821.41837.418441st or higher1851.718601873.7188Maximum189451903319145192Minimum193319431953196Mean19717.219816.419918.1200Standard deviation2018.762027.602039.89204Age of Building2055 years or less206N.A.20769.62086–10 years20927.7210More than 10 years2112.7212Maximum21311214Minimum2150216Mean2174.71218Standard deviation2192.41220Other characteristics221Accessibility to MTR station22262.122376.622444.9225Located in Cheung Sha Wan2268.42274.622812.8229BEAM Plus certification230Assessed23142.223277.82330234Provisional Platinum rating235023602370238Provisional Gold rating2398.724016.02410242Provisional Silver rating2432.52444.62450246Provisional Bronze rating24718.424833.92490250Unclassified25112.725223.42530254Registered but not assessed25512.125622.22570258259260===== PAGE 10 (score 10, numeric-paren 0, table-words 1, var-words 8) =====261262393263The effect of BEAM Plus certification on property price in Hong Kong264sampled residential flats are within 201–400 sq. 265ft. The highest flat in the sample is on the 45th 266floor whereas the lowest one is on the 3th floor. On 267average, the floor level of the residential units is 26817.2, and more than 65% of them are on the 20th 269floor or lower. In addition, 62.1% of all flats in the 270data sample are within 5-minute walking distance 271to the MTR station, and less than 10% of them are 272located in Cheung Sha Wan. Lastly, 53.3% of the 273flats in the sample are located in buildings which 274have been registered under BEAM Plus, and 42.2% 275of them in already-assessed buildings.276As for the pre-sale housing units and the occu-277pied housing units in the sample, the transaction 278price of the pre-sale flats is higher than that of the 279completed flats. Almost three quarters of all pre-280sale flats were sold at more than 4 million HKD, 281while only 42.9% of the occupied flats were trans-282acted at the same price range. By contrast, the av-283erage size of the pre-sale flats is smaller than that 284of the occupied flats. While close to 70% of pre-285sale flats in the data sample are 201–400 sq. ft., 286the percentage of occupied flats with similar floor 287area is less than 60%. Also, the mean floor level 288of the pre-sale flats is slightly lower than that of 289the occupied flats as well. The average age of the 290occupied flats is 4.7 years, with the oldest being 29111 years old and the newest only occupied for less 292than half a year. A much larger proportion of pre-293sale flats in the data sample are within 5-minute 294walking distance to the MTR station, compared 295to the occupied flats, while a larger percentage 296of occupied flats are located in Cheung Sha Wan. 297Regarding BEAM Plus certification, however, it 298should be noted that all of the BEAM-Plus-regis-299tered flats in the sample are pre-sale. 78% of these 300flats are located in buildings that have already 301been assessed under BEAM Plus, and the highest 302proportion of them are within buildings certified 303with a Provisional Bronze rating.304As this study concerning the price impact of 305BEAM Plus certification on residential flats in 306Hong Kong, relevant data which captures the 307structural attributes, neighborhood attributes, and 308BEAM Plus certification attributes is required. The 309data comes from three sources. The first source is 310the property transaction database compiled from 311the Economic Property Research Centre (EPRC), 312in which all property transactions recorded by the 313Hong Kong Land Registry since 1991 are avail-314able. In these records, information concerning the 315structural attributes and neighborhood attributes 316of the transacted properties, such as transaction 317price, floor area, floor level, name/age/address of 318the building in which the sampled housing flat 319is located, is provided. Nevertheless, it should be 320noted that, though critical, some elements of these 321transacted housing units are not included in the 322EPRC database, for instance the quality of build-323ing and design quality.324The second source of data is the Rating & Valua-325tion Department (RVD). As noted in the paragraph 326above, only the transaction price at the time when 327a housing unit was sold is included in the transac-328tion records within the EPRC database. However, 329if this raw price data is incorporated into the em-330pirical model, the property price trends over time 331would be overlooked, and the results would likely 332be distorted. To take this property price trend into 333consideration, the raw price data is needed to be 334adjusted with reference to Hong Kong’s residential 335housing price trends over time. Thus, the RVD’s 336property price index (SM4.2) is used for the com-337putation of the adjusted transaction price of the 338sampled housing units.339Most importantly, for information about the 340sampled buildings’ status with reference to BEAM 341Plus certification (or registration) we rely on two 342sources: 1) the BEAM Society and 2) the HKGBC. 343While the former generally provides technical in-344formation concerning BEAM Plus itself, the lat-345ter’s official website proffers other important in-346formation such as the most updated BEAM Plus 347certification/registration statistics by sector (and 348the status of registered buildings).349As for the methodology, the hedonic price mod-350el is used. The conceptual framework behind this 351model is derived from Rosen’s (1974) framework in 352which a variety of products are differentiated and 353their utility-bearing attributes valued. Within the 354context of housing, Malpezzi (2002) points out that 355the hedonic equations can decompose the value of 356housing into measurable prices and quantities, 357and that the resultant regression coefficients may 358be transferred into the implicit prices of various 359housing attributes.360Regarding the variables required for this in-361vestigation, the dependent variable in the model 362is the per square metre transaction price of the 363sampled housing units, adjusted with respect to 364the general property price indices when a housing 365unit was transacted. It should be noted that, the 366model follows a semi-log specification, in that the 367(adjusted) per square metre transaction price in 368natural log form (LnPPSM) is also deployed. The 369reason behind such a specification is twofold. First, 370the interpretation of the results is more straight-371forward, as the price premium generated from the 372373374===== PAGE 13 (score 45, numeric-paren 25, table-words 5, var-words 7) =====375376396377E. C. M. Hui et al.378ANOVA test results, which indicate that the null 379hypothesis is rejected at 5% level for all four mod-380els, and that the models are considered statisti-381cally significant.382For Model 1, the findings first illustrate that 383BEAM Plus certification is statistically significant 384(within 1% level) in explaining the transaction 385price of the sampled properties. The coefficient is 386positive, which means that a housing unit located 387in a BEAM-Plus-certified building is transacted 388at a price about 4.4% higher than a similar hous-389ing unit located in a non-registered building. This 390is consistent with the literature, in that green 391building certification generates additional price 392premium for housing units. Model 1 also reports 393a statistically significant (within 1% level) rela-394tionship between UNCL and property price per 395square metre. The negative coefficient suggests 396that, holding other variables constant, a housing 397unit located in a building with an “unclassified” 398rating under BEAM Plus is sold at a discount of 3995.9% when compared with similar housing units in 400non-registered buildings. This finding is consistent 401to those in recent studies on the EPC (Fuerst et al. 4022015, 2016). Viewing these two findings together, 403it can be said that there exists a 10.3% difference 404in price between a housing unit in a certified build-405ing and a similar flat in a registered building with 406an “unclassified” rating.407As for the structural attributes, the findings in-408dicate that a flat’s saleable area has a significant 409relationship with it’s (per square metre) transac-410tion price. The coefficient suggests that, for every 4111% increase in saleable floor area, a flat’s (adjust-412ed) transaction price per square metre would be 4130.005% lower. Another factor found to be nega-414tively related with transaction price is age. Using 415pre-sale flats as the base case, it is reported that 416the older the building in which a housing unit is 417located, the lower per square metre price it has. 418Specifically, the price per square metre of a hous-419ing unit in a building which has been completed 420for 5 years or less is 18.5% lower than that of a 421pre-sale housing unit. The discounts in price for 422housing units in buildings that are 6–10 years old 423and more than 10 years old are 23.8% and 32.4%, 424respectively. This is in line with the expectation 425in the relationship between property price and 426the age of residential buildings. By contrast, the 427floor level dummy variables are all positively re-428lated (significant at 1% level) in explaining the 429per square metre transaction price of the sampled 430housing units. Using flats on the 10th floor or be-431low as the base case, it is found that flats are sold 432at prices approximately 3% (11th to 20th floor), 9% 433(21st to 30th floor), 15.4% (31st to 40th floor), and 43419.9% (41st floor or above) higher. This is consist-435ent with the expected relationship between floor 436level and property price.437Lastly, for the neighborhood attributes, the 438findings reveal that better accessibility, as expect-439ed, results in higher property prices, as a residen-440tial unit’s proximity to MTR station (i.e. 5-minute 441walking distance) is found to incur a price premi-442um of 2.8% when compared to others, controlling 443for the effects of other variables. The NSSP vari-444able, likewise, is statistically significant within 1% 445level, in that a residential unit located in Cheung 446Sha Wan, is about 24.2% more costly than a sim-447ilar flat located in Sham Shui Po, holding other 448variables constant.449For Model 2, without UNCL as a control vari-450able, the resultant transaction price premiums/dis-451counts differ in all categories with the exception 452of floor area, even though their signs are identi-453cal. The findings show, first, that the coefficient 454for the CERTIFIED variable becomes noticeably 455higher, in that a housing unit in a BEAM-certified 456building is 6.2% more expensive than another flat 457with identical characteristics in a non-registered 458building.459Then, for the structural attributes, the nega-460tive price impact of a residential building’s age 461Table 7. The results from the hedonic pricing models462Model 1463Model 2464Constant46512.110**46612.062**467BEAM Plus certification attributes468CERTIFED4690.044**4700.062**471UNCL472–0.059**473Structural attributes474AREA475–0.005**476–0.005**4775 years or less478–0.185**479–0.157**4806 to 10 years481–0.238**482–0.203**483More than 10 years484–0.324**485–0.295**48611th to 20th floor4870.030**4880.031**48921st to 30th floor4900.090**4910.091**49231st to 40th floor4930.154**4940.153**49541st floor or above4960.199**4970.188**498Neighborhood attributes499MTR5000.028**5010.051**502NSSP5030.242**5040.250**505N506646507Adjusted R-square5080.8005090.796510F-statistic511216.095512229.644513Note: ** denotes significance at 1% level; and * at 5% level.514515516===== PAGE 14 (score 8, numeric-paren 0, table-words 0, var-words 8) =====517518397519The effect of BEAM Plus certification on property price in Hong Kong520is not as prominent as it is in Model 1. Using 521pre-sale housing units as the base case, the price 522discounts for flats located in older buildings are 52315.7% (5 years old or less), 20.3% (6–10 years old), 524and 29.5% (more than 10 years old), respectively. 525By contrast, using flats on the 10th floor or below 526as the base case, flats on lower floor levels (11th–52730th floor) carry a slightly higher price premium 528compared to Model 1 (3.1–9.1%), whereas those on 529higher floor levels (31st floor or above) result in 530slightly lower price premium (15.3–18.8%).531Lastly, for the neighborhood attributes, the re-532sultant price premium for a flat with accessibility 533to either Sham Shui Po MTR station or Cheung 534Sha Wan MTR station, at 5.1%, is noticeably 535higher than that in Model 1 (2.8%). Also, a flat 536located in Cheung Sha Wan is, on average, sold at 537a per square metre price about 25% higher than a 538similar flat in Sham Shui Po, compared to 24.2% 539in Model 1.5406. CONCLUSION AND POLICY 541IMPLICATIONS542This study has evaluated the price effect of BEAM 543Plus certification on flats in sixteen residential de-544velopments within the Sham Shui Po – Cheung 545Sha Wan district. The findings first show that the 546prices of flats in BEAM Plus-certified buildings are 5474.4% higher than those in non-registered build-548ings, and that flats in buildings with an “unclas-549sified” rating under BEAM Plus are transacted at 550a discount of 5.9%. To put it differently, the price 551difference between a BEAM-Plus-certified flat and 552a similar flat in an “unclassified” building amounts 553to 10.3%. Nevertheless, this is only the case when 554homebuyers are informed of the actual status of a 555residential development with respect to its BEAM 556Plus assessment (including the “unclassified” rat-557ing). For less-informed homebuyers, however, 558since property developers only disclose information 559about a development’s participation in BEAM Plus 560assessment when it is certified, it is very likely 561that these buyers would mistake an “unclassified” 562building for a non-registered building. Then, the 563price premium of BEAM Plus certification becomes 564higher (6.2%). In other words, the non-disclosure 565of BEAM Plus assessment results of “unclassified” 566buildings on the part of the developers would lead 567to over-valuation of price premium.568Regardless, considering the additional cost (2%) 569required for the construction of green buildings 570(Kats 2003, 2006; Morrison Hershfield 2005; Ber-571ry 2007), the price premium as a result of BEAM 572Plus certification suggests that developers may 573find it profitable, or at least financially viable, to 574“go green” in response to the market demand for 575green living space. From another perspective, the 576price premium obtained via BEAM Plus certifica-577tion (4.4–6.2%, depending on model specifications) 578are generally on par with those found in the resi-579dential sector (Chegut et al. 2010; Shimizu 2010; 580Deng et al. 2012; Hyland et al. 2013; Jayantha, 581Man 2013; Brounen, Kok 2011; European Com-582mission 2013; Heinzle et al. 2013; Deng, Wu 2014; 583Kahn, Kok 2014; Fuerst et al. 2015, 2016; Bond, 584Devine 2016; Bruegge et al. 2016), but much lower 585than those found in the office sector (Miller et al. 5862008; Fuerst, McAllister 2008, 2009, 2011; Eich-587holtz et al. 2008, 2009, 2010, 2013; Murray 2008; 588Pivo, Fisher 2009; Wiley et al. 2010; Reichardt 589et al. 2012; Chegut et al. 2014; Kok, Jennen 2012). 590While some of the benefits are applicable to own-591ers of residential space and of office space, such as 592higher rents (or returns), lower holding costs, and 593lower risk premium (see Fuerst, McAllister 2008, 5942011; Aroul, Hansz 2012), the reason behind the 595disparities in price premium between the two sec-596tors can be attributed to the differences in the tan-597gible and intangible benefits associated with green 598living space and green office space. For the tangi-599ble benefits, approximately 55% of all electricity 600consumed in Hong Kong’s office sector in 2014 was 601on space conditioning, compared to around 35% in 602the residential sector (Fig. 3). Therefore, the daily 603operating costs, in terms of electricity consump-604tion, for office use are much larger than those for 605residential use. Operating in a building with bet-606ter environmental performance would reduce the 607consumption of electricity, thus lowering operating 608costs for businesses. The improvement in energy 609costs for occupiers of living space in green build-610ings, in comparison, would not be as remarkable. 611For the intangible benefits, while the manufactur-612ing sector can illustrate its commitment to sus-613tainability by producing environmental-friendly 614products, such commitment is not as noticeable for 615non-manufacturing sectors. Therefore, one way to 616showcase this commitment, and thus boost corpo-617rate image, is to move their offices to green build-618ings. By contrast, even though buyers of green 619living space may take the “feel good” factor into 620consideration, this factor does not translate into 621business opportunities that an enhanced corporate 622image does.623The findings in this study have some implica-624tions regarding the GFA concession policy. This 625policy, on paper, provides some extra incentives 626