Competing or Complementary Labels Appendix

Appendix A: Formal Statement of Hypotheses A project chooses to certify using LEED, GBEL, neither or both. Let L∈{0,1} and G∈{0,1} denote the decision of a focal project to adopt LEED or GBEL respectively. Let S represent the stock of GBEL certified buildings in the same market as the focal project. We can write the net benefits (payoffs) from certification as where the function B represents gross benefits of certification and C the associated costs. Using this notation, we can precisely state the hypotheses as follows: Proposition 2 (Supplier Spillovers) is equivalent to assuming that 𝜋(L, G; S) is super-modular in (L,S) through the cost function Proposition 3 (Adopter Spillovers) is equivalent to assuming that 𝜋(L, G; S) is super-modular in (L,G) Proposition 4 (User Spillovers) is equivalent to assuming that 𝜋(L, G; S) is super-modular in (L,S) through the gross benefits function The text describes in greater detail a number of mechanisms that could generate these types of payoffs. Proposition 1 is simply a generalization of Propositions 2 through 4, noting that any of those different mechanisms could generate complementarities in the adoption of related voluntary certification schemes. Appendix B: Data and Variable Construction To compile a database of green buildings in China, we hand collect and translate GBEL certification data from MOHURD, and obtain the LEED certification and LEED AP data directly from the USGBC. We merge the two sets of data to obtain a comprehensive dataset of green building adoptions in China. We also collect demographic information from China Census. Our analysis use data on 657 cities with green buildings in China from 2005 to 2015. The central government sets more stringent energy efficiency targets under the 13th Five Year Plan (2016 – 2020) and some cities/provinces require all new buildings to achieve the green building standard so data after 2016 might not be a good fit for the research design. LEED Certification: we obtained annual LEED certification data from the USGBC directly. We have the building level data including building name, building owner, year of certification, building type, certification type, located city, and certification rating level. Information of around 12% of LEED certified buildings is confidential. As we don’t have some basic information such as the located city, building type, and rating level of these confidential buildings, they are not included in the empirical analysis. GBEL Certification: we collect annual GBEL certification data from 2005 to 2015 from MOHURD20. We have the building level data including the building name, year of certification, building type, certification type, located province, and certification rating level. We translate all information from Chinese to English and search the cities where each project is located based on building names so that we can have a city-level panel data analysis. This is comparable with the LEED Certification data discussed above. Green building policies and financial subsidies: we gather the national, provincial, and city level green building policy and financial subsidies from MOHURD and provincial government websites. A total of 7 provincial-level administrative units (China has 34 provincial-level administrative units: 23 provinces, 4 municipalities, 5 autonomous regions and 2 special administrative regions) in China have provided their own financial subsidies on GBEL adoption during the 12th FYP period. Four cities, including Luoyang, Xi’an, Qingdao, and Wuhan also provide city level financial subsidies. Because the provincial and city level financial subsidies do not reflect any firm’s strategic decision, such “treatment” offers exogenous variation in firms’ decision in adopting green building practice. Construction activity: we collect yearly area of built district in each city from MOHURD. Demographics: for each city in the analysis, we obtain urban district population in different cities from MOHURD and China Census. LEED AP: we collect the annual LEED AP data from the USGBC directly. We have the number of new LEED APs in China by city and by year from 2005 to 2015. Environmental preferences: we measure city level environmental sustainability by “China Green LowCarbon City Index” of 114 cities co-assessed by Lawrence Berkeley National Laboratory, University of San Francisco, Innovative Green Development Program, and Energy Foundation China (Ohshita et al, 2017). It not only provides the total green low-carbon city score of each city, but also have scores on economy, energy/power, transport, industry, buildings, environment/land, policy/outreach of these cities. The new combined data are not without drawbacks. First, we do not have information on GBEL registration data, which may better measure firms’ intentions to build green, compared with certification data we obtain and use in the analysis. Second, we don’t know detailed information of buildings that are not certified by either LEED or GBEL in different cities, and don’t know the proportions of green certified buildings in different cities. As such, our key empirical analysis is at the city level instead of the building level. In order to analyze multiple certification adoption, we match the building level data from MOHURD and USGBC. Both data sets are translated from Chinese to English (we translate the GBEL certification data while USGBC or LEED building applicants translate the LEED certified building names). The translation, which can be produced by pronunciation or meaning or neither, makes the building name matching very challenging. For instance, “北辰” can be translated into “Beichen” (pronunciation) or “North Star”(ancient Chinese meaning). Also, some building developer may assign an English name to their buildings with almost no relevance to the Chinese name, but the translator may not be aware this English name. For instance, a building developed by Zijin Group is called “W Square”, but the Chinese name of the building is actually “Zhonghang Zijin.” We have conducted several rounds of checks to minimize the building name discrepancies between the two sets of data, but due to the potential information lost in translation explained above, the total number of buildings adopting multiple certifications analyzed might be slightly underestimated. 20 http://www.cngb.org.cn/cms/view/index.action?sid=402888b44f81b20f014f81dd5b21000c Appendix C: Supplemental Tables Table C1. Robustness Check 1 - Sample with at least one green buildings First Stage lnGBEL lnGBEL SUBSIDY SUBYEAR lnGBEL 0.41 (0.11) 0.24 (0.06) Control Variables lnArea Reduced Form lnLEED 0.12 (0.05) lnLEED Spillover Effect OLS lnLEED IV lnLEED 0.21 (0.04) 0.29 (0.10) 0.07 (0.03) 0.22 (0.11) 0.22 (0.12) 0.01 (0.05) 0.01 (0.05) -0.04 (0.04) -0.06 (0.04) Year Dummy Yes Yes Yes Yes Yes Yes Observations 2466 2466 2466 2466 2466 2466 lnPop City Fixed Effects R-Squared -0.02 (0.03) -0.02 (0.03) Yes Yes 0.48 0.48 0.01 (0.01) Yes 0.09 Notes: Unit of analysis is a city-year. Robust standard errors clustered by city in parentheses. 0.01 (0.01) Yes 0.08 0.01 (0.01) Yes 0.25 0.01 (0.01) Yes Table C2. Robustness Check 2 – Green Building Counts instead of Logs First Stage GBEL Count lnGBEL SUBSIDY SUBYEAR 1.48 (0.53) Control Variables lnArea Reduced Form GBEL Count 1.41 (0.50) LEED Count 0.24 (0.17) LEED Count Spillover Effect OLS LEED Count IV LEED Count 0.15 (0.07) 0.12 (0.09) 0.18 (0.14) 0.06 (0.01) 0.06 (0.01) 0.01 (0.00) 0.01 (0.00) 0.00 (0.00) 0.00 (0.00) Year Dummy Yes Yes Yes Yes Yes Yes Observations 6822 6822 6822 6822 6822 6179 lnPop City Fixed Effects R-Squared 0.01 (0.00) Yes 0.30 0.01 (0.00) Yes 0.31 0.00 (0.00) Yes 0.10 Notes: Unit of analysis is a city-year. Robust standard errors clustered by city in parentheses. 0.00 (0.00) Yes 0.10 0.00 (0.00) Yes 0.31 0.01 (0.00) Yes Table C3. Robustness Check 3 – Different Control Variables First Stage lnGBEL lnGBEL SUBSIDY SUBYEAR Reduced Form lnGBEL 0.30 (0.07) 0.18 (0.04) lnLEED 0.07 (0.03) lnLEED Spillover Effect OLS lnLEED IV lnLEED 0.19 (0.03) 0.24 (0.07) 0.04 (0.02) Year Dummy Yes Yes Yes Yes Yes Yes Observations 6889 6889 6889 6889 6889 6889 City Fixed Effects R-Squared Yes Yes 0.19 0.19 Yes 0.04 Notes: Unit of analysis is a city-year. Robust standard errors clustered by city in parentheses. Yes 0.04 Yes 0.24 Yes Table C4. Robustness Check 4 – GBEL Certification and LEED Registration First Stage lnGBEL lnGBEL SUBSIDY SUBYEAR 0.28 (0.07) Control Variables lnArea lnGBEL 0.17 (0.04) lnLEEDReg 0.07 (0.03) lnLEEDReg Spillover Effect OLS lnLEEDReg IV lnLEEDReg 0.22 (0.03) 0.21 (0.08) 0.03 (0.01) -0.02 (0.01) 0.22 (0.06) -0.02 (0.01) 0.06 (0.04) 0.06 (0.04) 0.01 (0.02) 0.02 (0.03) Year Dummy Yes Yes Yes Yes Yes Yes Observations 6822 6822 6822 6822 6822 6179 lnPop City Fixed Effects R-Squared 0.22 (0.06) Reduced Form Yes 0.20 Yes 0.20 -0.01 (0.01) Yes 0.04 Notes: Unit of analysis is a city-year. Robust standard errors clustered by city in parentheses. -0.01 (0.01) Yes 0.04 0.00 (0.01) Yes 0.20 0.00 (0.01) Yes