What Determines Chinese Firms Decision on Implementing Voluntary Environmental Schemes?

Transcription

1 Journal of Servce Scence and Management, 2011, 4, do: /jssm Publshed Onlne September 2011 ( What Determnes Chnese Frms Decson on Implementng Voluntary Envronmental Schemes? Jun-Y Shen 1, Xang-Dong Qn 2 1 Faculty of Internatonal Studes, Hroshma Cty Unversty, Hroshma, Japan; 2 School of Economcs, Anta College of Economcs and Management, Shangha Jao Tong Unversty, Shangha, Chna. Emal: shen@ntl.hroshma-cu.ac.jp, xdqn@sjtu.edu.cn Receved May 4 th, 2011; revsed July 6 th, 2011; accepted July 16 th, ABSTRACT Understandng why frms choose to mplement voluntary envronmental schemes n a large polluted country lke Chna s mportant for both envronmental economsts and polcy makers. In ths paper, we utlze unque plant-level survey data of 270 Chnese frms n manufacturng ndustry to dentfy the key determnants of ther decsons on certfyng ISO envronmental management standard and the Chnese Envronmental Label. The emprcal results exhbt that whle there are a number of factors (e.g., ownershp, frm sze, target market, and the number of rvals) havng smlar effects on the certfcaton decsons between the two examned programs, the unque factors that only affect the decson of certfyng one program (.e., ISO or the Chnese Envronmental Label) are also observed. Keywords: Voluntary Envronmental Scheme, ISO 14001, Eco-Label, Chnese Frm 1. Introducton In the past decade, the trend towards promotng voluntary acton and polluton preventon as opposed to mandatory (command-and-control) envronmental regulatons that prescrbe quantty lmts on pollutants had ncreased as more governments worldwde faced and contnue to face lmted envronmental enforcement budgets [1]. The emergng voluntary approach to polluton abatement becomes more and more popular n recent years. Ths s because of a growng number of busness-ntated actons to change corporate culture and management practces through the ntroducton of envronmental label or eco-label programs and nternatonal envronmental management system certfcaton programs such as the Internatonal Standards Organzaton (ISO) programs [2]. Eco-labels have been used for over twenty years to provde consumers wth nformaton about a product whch s characterzed by mproved envronmental performance and effcency compared wth smlar products wthout those labels. The man purpose of promotng eco-label programs s to avod the possble asymmetrc nformaton problem between producers and consumers, because some envronmentally frendly products normally have unobservable characterstcs [3]. In comparson to ecolabel programs, ISO programs concern the way a frm goes about ts producton, and notdrectly the results of ths producton. In other words, they concern processes, and not products at least, not drectly [1]. Among them, the ISO s an nternatonal, voluntary standard for envronmental management promoted by the ISO, whch formally ncludes fve steps n an envronmental management system as envronmental polcy, plannng, mplementaton and operaton, checkng and operaton, and management revew. 1 In the lterature, a number of studes have been undertaken on the determnants of frms envronmental performance and frms voluntary pro-envronmental acton [5-18]. Among these studes, [12] employed panel data models to study how envronmental nnovaton by US manufacturng ndustres responds to changes n polluton abatement expendtures and regulatory enforcement durng the perod of 1983 to They found that envronmental nnovaton responds to ncreases n polluton abatement expendtures, however, ncreased montorng and enforcement actvtes related to exstng envronmental regulatons do not provde any addtonal ncentve to nnovate. [9] examned data on complance wth envronmental regulatons wthn the manufacturng sector n Mexco and found that the probablty of complyng depends on the knd of management practces of the 1 For more detaled ssues on the ISO management standard, see [4].

2 What Determnes Chnese Frms Decson on Implementng Voluntary Envronmental Schemes? 381 frm and the level of envronmental tranng. [14] nvestgated whether frms characterstcs nfluence ther decsons to jon the Envronmental Protecton Agency s voluntary Green Lghts program. They reported that contrary to conventonal theory, the characterstcs of frms do affect ther decsons to jon Green Lghts and commt to a program of nvestments n lghtng effcency. [11] examned emprcally the determnants that led large Japanese manufacturers to voluntary envronmental commtment and found that the costs and benefts of voluntary actons to enhance or protect the envronment, the capacty to act, as well as the envronmental values, belefs, and atttudes of managers are sgnfcant determnants of voluntary envronmental commtment. In addton to the above researches on the factors nfluencng frms envronmental performance and ther voluntary pro-envronmental acton, we fnd two studes n the lterature pontng to the mpact of envronmental performance on frm performance. [19] conducted a case study to examne the mpact of envronmental ratng of large pulp and paper, auto, and chlor alkal frms on ther stock prces and found that the market generally penalzes envronmentally unfrendly behavor. [20] conducted statc and dynamc panel data analyss of the mpact of envronmental performance on the frm s fnancal performance. They showed that envronmental performance has a neutral mpact on the frm s fnancal performance, whch s consstent wth theoretcal work suggestng that frms nvest n envronmental ntatves untl the pont where the margnal cost of such nvestments equals the margnal beneft. In recent years, Chna s envronmental problem, whch s consdered as a byproduct of ts rapd economc growth, s becomng a well-known ssue and recevng more and more attentons from both economsts and envronmental specalsts around the world. Prevous studes n the envronmental economcs lterature related to envronmental ssues of Chnese frms pad a lot of attentons to the fnancal ncentves and endogenous enforcement n Chna s polluton levy system (e.g., [15]), ndustral ownershp and envronmental performance n Chna (e.g., [17]), barganng power of Chnese factores n enforcement of polluton regulaton (e.g., [13]), relatons among nspectons, polluton prces and Chnese envronmental performance (e.g., [8]), and Chnese frms technology development and energy productvty (e.g., [16]). However, the ssue on the factors nfluencng Chnese frms practce of voluntary envronmental schemes (e.g., the ISO and eco-label) has not been fully studed. For ths purpose, a frm-level survey was conducted n 2008 to obtan a better understandng of a frm s motvatons, ts decson-makng procedures and ts characterstcs vs-à-vs ts decson on jonng a voluntary envronmental program. We consder two envronmental programs here, one s ISO certfcaton and the other s Chna Envronmental Label (herenafter referred to as eco-label). Consequently, ths paper tres to analyze the determnants that lead Chnese frms to mplement ISO and/or eco-label certfcatons. More specfcally, we provde emprcal results regardng what determnes a frm s choce as to: 1) whether or not t should ntroduce a certfed envronmental management system (.e., ISO 14001); 2) whether or not t should mplement a certfcaton of eco-label (.e., Chna Envronmental Label). The paper s organzed as follows. Secton 2 presents our data and Secton 3 dscusses our emprcal methodology. Our emprcal fndngs are reported and dscussed n Secton 4. The fnal secton hghlghts polcy mplcatons of the results and provdes suggestons for future research. 2. Data To nvestgate the determnants of Chnese frms choce of adoptng eco-label and ISO standard, we conducted a frm-level survey n Chna n Detaled face-to-face ntervews were conducted at 270 frms, whch were chosen to represent Chna s manufacturng ndustry n a set of categores defned by sector, sze class and ownershp structure. The majorty of the ntervews were focused on Yangtze Delta Area, whch conssts of Shangha, Zhejang Provnce, Jangsu Provnce, and Anhu Provnce, to represent ts strong economc performance n Chna n recent years. 2 As shown n Table 1, the number of frms from ths area s 196, makng up 72.59% of the sample. In our vew, the sample s well balanced among 13 sector categores based on the study of [16]. They are 37 frms n the chemcals sector, 30 n the food and beverage sector, 57 n the machnery, equpment and nstruments sector, 21 n the metal processng and products sector, 16 n the nonmetal products sector, 19 n the textle, apparel and leather products sector, 15 n the tmber, furnture, and paper products sector, 14 n the medcal products sector, 10 n the rubber and plastc products sector, 9 n the electrc power sector, 7 n the automoble and related products sector, 4 n the petroleum processng and cokng sector, and 31 n other sectors of manufacturng ndustry. The frms are also evenly dstrbuted along the sze scale, wth roughly smlar numbers n large, medum, and small classes. Sze classes are defned by employment ranges, wth small frms employng employees, medum about employees and large more than 1000 employees. About 40% n 2 The GDP share of the Yangtze Delta Area over the whole country s about 25.45% n 2006, whle the populaton share s about 15.56% [21].

3 382 What Determnes Chnese Frms Decson on Implementng Voluntary Envronmental Schemes? Table 1. Summary of the sample. Characterstcs n % Characterstcs n % Sector Ownershp Chemcals State-owned Electrc power Collectve-owned Food and beverage Prvate-owned Machnery and nstruments Foregn Medcal products Jont-venture Metal processng and products Automoble and related products Has eco-labeled products Nonmetal products Yes Petroleum processng/cokng No Rubber and plastc products Textles and leather products Has ISO certfcaton Tmber, furnture, and paper Yes Other sectors No Area Total revenue n 2007 Shangha <10 mllon RMB Zhejang mllon RMB Jangsu mllon RMB Anhu mllon RMB Other provnces mllon RMB mllon RMB Lsted frm on the stock market mllon RMB Yes >1000 mllon RMB No No answer Number of the employees Number of the rvals < < > > Total Total the sample s prvate-owned, whle the numbers of stateowned, foregn-owned, and jont-venture frms are roughly the same. In addton, 23.33% of frms are publcly traded n the two stock markets n Chna (Shangha Stock Exchange and Shenzhen Stock Exchange). 259 frms answered ther annual sales amount n the year of 2007 n our desgnated ranges, wth roughly. smlar share n each category. 3 In the ntervew, we asked the respondent whether the frm has any eco-labeled products and whether the frm has the certfcaton of ISO standard. 69 frms 3 The purpose to ask frms to report ther annual sales amount n ranges nstead of n amounts s to avod possble complants. reported that they had at least one eco-labeled product and 77 frms have already certfed the ISO The number of frms wth both certfcaton s 23, accountng for 8.52% of the sample. As for the reasons of the eco-label certfcaton gven by those frms wth eco-labeled products, 73.91% (51 frms out of 69) concurred that products wth the ecolabel certfcaton would be more easly accepted by consumers, whle 40.58% admtted that they dd so because of the government regulaton. In addton, there were 52.17% of the frms wth the eco-label certfcaton reported that they regarded eco-labelng as a strategy for promotng the frm s prestge, and 57.97% answered that

4 What Determnes Chnese Frms Decson on Implementng Voluntary Envronmental Schemes? 383 obtanng an eco-label s to demonstrate the effort for envronmental protecton made by the frms. The last reason that the certfcaton fee for eco-labelng s not so hgh was cted by 15.94% of the sample. Wth respect to the reasons of ISO certfcaton, 60 of the 77 frms reported that ther purpose was to reduce producton costs, 63 of the 77 frms admtted that they wanted to reduce the envronmental load caused by ther producton, and 63 of the 77 frms stated that ther purpose was to rase ther staffs envronmental conscousness. Addtonally, 54.55% of the frms wth the ISO certfcaton reported that mplementng ISO management system would help them communcate wth outsde more easly, 75.32% answered that they dd so to ncrease the frm s credblty, and 62.34% sad that ther purpose was to upgrade the frm s prestge. Fnally, the number of the frms that cted other reasons (e.g., as a combnaton wth other management systems, etc.) was 11, makng up 14.29% of the ISO certfed frms. 3. Methodology Ths paper seeks to dentfy the factors that nduce the certfcatons of eco-label and/or ISO The frm s decson to choose to certfy s descrbed by the followng latent varable models. * Ecolabel ISO X (1) * X e (2) where Ecolabel and ISO denote whether a frm has eco-labeled products and whether a frm has ISO certfcaton, respectvely. X represents a matrx of explanatory varables capturng the factors that may affect the certfcaton decson of a frm (see more detaled descrptons of these factors n Table 2). and are coeffcents matrx of X. 1, 2,3, 4,, 270 represents each frm, and and e represent the error terms n Equatons (1) and (2), respectvely. The varables marked wth astersks n the equatons are latent varables. They are lnked to the observed varables n the followng way. Ecolabel 1 a frm has eco-labeled products 0 otherwse (3) 1 a frm has ISO certfcaton ISO otherwse (4) The above model can be estmated usng a bnary probt model f we assume the error terms n Equatons (1) and (2) follow the standard normal dstrbuton. It s mportant to note that the parameters of the model, lke those of any nonlnear regresson model, are not the margnal effects we are accustomed to analyzng. The computaton of the margnal effect of a contnuous varable n a probt model s performed n the followng way. E[Ecolabel / X] ( X ) (5) X E[ISO 14001/ X] ( X ) (6) X where () s the standard normal densty. In ths case, the margnal effects are usually evaluated at the sample means of the data. Furthermore, another complcaton for computng the margnal effect n a bnary probt model arses because X wll often nclude dummy varables, as the case n ths study. The computaton of the margnal effect (ME) of a dummy varable can be done n the followng way. ME P(Ecolabel 1 x ) P(Ecolabel 1 x ) (7) ( d 1) ( d 0) ME P(ISO x( d 1) ) P(ISO x( d 0) ) (8) where d s a dummy varable and x ( d ) denotes the means of all the other varables n the model. 4. Results Tables 3 and 4 present results from a probt model of the factors assocated wth eco-label and ISO certfcatons. The robust standard errors of the coeffcents are reported n the tables. In addton, we also calculate the margnal effects accordng to Equatons (5)-(8) and report the results n Table 5. The detaled descrptons of the varables are provded n Table Eco-label Certfcaton Table 3 ndcates the exstence of sgnfcant correlatons between the reasons of eco-label certfcaton admtted by the frms and the certfcaton decson n all cases (models (1-4)). It s mportant to note that the presented reasons provde dfferent ncentves. For example, Accepted_consumer and Promoton_strategy are more lkely related to frm s economc ncentve, whle Envronment_protecton seems to reflect that frms have a strong motvaton to take a pro-envronment approach. Whatever the dfferences of these ncentves are, the estmated postve sgns of these varables suggest that relatve to the omtted reason (.e., the certfcaton fee for eco-labelng s not so hgh), a frm motvated by any of these reasons s more lkely to choose eco-label certfcaton. Concernng the factor of ownershp, Foregn_ owned and Jont_venture have sgnfcantly postve sgns n all cases, suggestng that foregn captal partcpaton does have an mportant effect on the frm s envronmental certfcaton decson. Ths result may attest

5 384 What Determnes Chnese Frms Decson on Implementng Voluntary Envronmental Schemes? Table 2. Defnton of the varables. Varables Dependent varables Ecolabel ISO Explanatory varables Accepted_consumer Government_regulaton Promoton_strategy Envronment_protecton Cost_reducton Envronment_load Staff_conscousness Promote_communcaton Credblty_ncreasng Prestge_upgradng Foregn_owned Jont_venture State_owned Prvate_owned Other_enterprses Indvdual_consumer Government Lsted_enterprse Rvals Foregn_market Frm_sze Chemcals Electrc power Food & beverage Machnery Medcal Metal Automoble Nonmetal Petroleum Rubber & plastc Textles & apparel & leather Tmber & furnture & paper Descrpton = 1 f there s eco-labeled product n the frm = 1 f the frm has ISO seres certfcaton = 1 f the frm admts eco-labeled products are easly accepted by consumers = 1 f the frm admts eco-labelng s based on the government s regulaton = 1 f the frm regards eco-labelng as a strategy for promoton = 1 f the frm admts eco-labelng s an effort for envronmental protecton = 1 f the frm admts ISO certfcaton reduces producton costs = 1 f the frm admts ISO certfcaton reduces envronmental load caused = 1 f the frm admts ISO certfcaton rases staffs envronmental concern = 1 f the frm admts ISO certfcaton promotes communcatons wth outsde = 1 f the frm admts ISO certfcaton ncreases the frm s credblty = 1 f the frm admts ISO certfcaton upgrades the frm s prestge = 1 f the frm s owned by foregn captals = 1 f the frm s a jont-venture enterprse = 1 f the frm s a state owned enterprse = 1 f the frm s owned by prvate sectors = 1 f the man clent of the frm s other enterprses = 1 f the man clent of the frm s ndvdual consumers =1 f the man clent of the frm s government = 1 f the frm s a lsted enterprse on the stock exchange market Numbers of the rvals reported by the frm = 1 f the frm manly sells ts products overseas Logarthm of the number of employees the frm has = 1 f the frm belongs to chemcal sector = 1 f the frm belongs to electrc power sector = 1 f the frm belongs to food and beverage sector = 1 f the frm belongs to machnery, equpment and nstruments sector = 1 f the frm belongs to medcal products sector = 1 f the frm belongs to metal processng and products sector = 1 f the frm belongs to automoble and related products sector =1 f the frm belongs to nonmetal products sector = 1 f the frm belongs to petroleum processng and cokng sector = 1 f the frm belongs to rubber and plastc products sector = 1 f the frm belongs to textles, apparel and leather products sector = 1 f the frm belongs to tmber, furnture and paper products sector that foregn owners are more eager to ncrease envronmental protecton ntatves to secure to prevent dscrmnaton or ncrease ther legtmacy n the eyes of these authortes. 4 In addton, the results show that the characterstcs of frms clent (.e, other enterprses, ndvdual consumers, and government) do not affect ther decsons on certfyng eco-label. Other frm characterstcs appear to have a strong mpact as well, although the factor that a frm s lsted on the stock exchange s not sgnfcant. 5 As expected, the

6 What Determnes Chnese Frms Decson on Implementng Voluntary Envronmental Schemes? 385 Table 3. Factors that affect the eco-label certfcaton (dependent varable: Ecolabel). (1) (2) (3) (4) Constant 1.933(0.177) *** 2.879(0.454) *** 5.538(0.976) *** 8.038(1.703) *** Reasons of certfyng eco-label Accepted_consumer 1.399(0.343) *** 1.498(0.381) *** 1.978(0.467) *** 2.470(0.755) *** Government_regulaton 1.572(0.520) *** 1.942(0.599) *** 1.904(0.849) ** 1.672(0.793) ** Promoton_strategy 0.698(0.389) * 1.001(0.438) ** 1.531(0.698) ** 1.754(0.869) ** Envronment_protecton 0.786(0.364) ** 0.705(0.392) * 1.303(0.658) ** 1.248(0.599) ** Ownershp Foregn_owned (0.482) *** 1.607(0.679) ** 1.787(0.799) ** Jont_venture (0.484) ** 2.283(0.692) *** 3.130(0.870) *** State_owned (0.554) * 0.985(0.705) 1.239(0.897) Prvate_owned (0.774) 1.434(0.767) * 1.155(0.918) Clents Other_enterprses (0.373) 0.736(0.591) Indvdual_consumer (0.351) 0.015(0.398) Government (0.432) * 0.264(0.658) Other frm characterstcs Lsted_enterprse (0.419) 0.319(0.496) Rvals (0.388) ** 1.680(0.676) ** Foregn_market (0.473) *** 1.139(0.533) ** Frm_sze (0.087) *** 0.435(0.124) *** Sector Chemcals (0.835) Electrc power (1.176) Food & beverage (0.933) ** Machnery (0.640) Medcal (0.738) Metal (0.822) * Automoble (0.856) ** Nonmetal (1.159) * Petroleum (1.398) ** Rubber & plastc (0.887) ** Textles & apparel & leather (0.539) Tmber & furnture & paper (0.719) Observatons Log lkelhood Pseudo R Notes: ***, ** and * denote statstcal sgnfcance at 1%, 5% and 10% levels, respectvely. Numbers n parentheses are robust standard errors. number of rvals s postvely correlated wth the certfcaton decson, ndcatng that the more brusng the competton a frm faces, the more lkely t certfes ecolabel to dstngush ts product from others. Foregn_ market s postve and sgnfcant, suggestng that exportorented frms are more nclned to have eco-label certfcaton. Accordng to [11], ths may occur because foregn consumers tend to be less able to montor the per-

7 386 What Determnes Chnese Frms Decson on Implementng Voluntary Envronmental Schemes? Table 4. Factors that affect the ISO seres certfcaton (dependent varable: ISO14001) (1) (2) (3) (4) Constant 1.628(0.147) *** 2.081(0.289) *** 3.778(0.770) *** 4.424(1.054) *** Reasons of certfyng ISO Cost_reducton 1.138(0.434) *** 1.295(0.459) *** 1.949(0.618) *** 2.737(0.802) *** Envronment_load 1.616(0.384) *** 1.512(0.413) *** 1.438(0.427) *** 3.414(0.748) *** Staff_conscousness 0.426(0.412) 0.331(0.451) 0.377(0.602) 0.626(0.943) Promote_communcaton 0.890(0.597) 1.953(0.618) 1.337(0.780) 2.151(0.983) Credblty_ncreasng 0.896(0.404) ** 0.972(0.411) ** 1.273(0.685) * 2.519(1.151) ** Prestge_upgradng 1.251(0.651) * 1.368(0.707) * 1.493(0.769) * 1.501(0.765) ** Ownershp Foregn_owned (0.372) *** 1.513(0.504) *** 2.726(0.686) *** Jont_venture (0.358) * 0.845(0.480) * 1.436(0.790) ** State_owned (0.449) 0.321(0.716) 0.748(0.785) Prvate_owned (0.651) 0.509(0.758) 1.055(1.160) Clents Other_enterprses (0.515) 1.044(0.603) * Indvdual_consumer (0.386) *** 2.079(0.503) *** Government (0.444) *** 1.739(0.516) *** Other frm characterstcs Lsted_enterprse (0.327) 0.109(0.426) Rvals (0.372) *** 2.524(0.509) *** Foregn_market (0.348) * 1.647(0.520) *** Frm_sze (0.111) *** 0.259(0.131) ** Sector Chemcals (1.052) *** Electrc power (1.810) *** Food & beverage (1.206) Machnery (0.641) Medcal (1.312) * Metal (0.609) Automoble (1.012) Nonmetal (2.054) Petroleum (0.824) *** Rubber & plastc (1.287) *** Textles & apparel & leather (0.525) Tmber & furnture & paper (1.265) Observatons Log lkelhood Pseudo R Notes: ***, ** and * denote statstcal sgnfcance at 1%, 5% and 10% levels, respectvely. Numbers n parentheses are robust standard errors. formance of the faclty or frm, as a result, more vsble sgns of envronmental commtment such as havng a certfed eco-label may legtmze ther reason for dong busness wth the frm. In addton, there s a sgnfcantly

8 What Determnes Chnese Frms Decson on Implementng Voluntary Envronmental Schemes? 387 Table 5. Margnal effects of the factors on the eco-label and ISO seres certfcatons. Eco-label certfcaton ISO14001 certfcaton Reasons of certfyng eco-label Accepted_consumer *** - Government_regulaton *** - Promoton_strategy * - Envronment_protecton ** - Reasons of certfyng ISO Cost_reducton *** Envronment_load *** Staff_conscousness Promote_communcaton Credblty_ncreasng ** Prestge_upgradng * Ownershp Foregn_owned ** *** Jont_venture ** * State_owned Prvate_owned Clents Other_enterprses * Indvdual_consumer *** Government *** Other frm characterstcs Lsted_enterprse Rvals ** *** Foregn_market ** ** Frm_sze *** ** Sector Chemcals ** Electrc power * Food & beverage ** Machnery Medcal * Metal * Automoble * Nonmetal ** Petroleum ** ** Rubber & plastc * ** Textles & apparel & leather Tmber & furnture & paper Notes: ***, ** and * denote statstcal sgnfcance at 1%, 5% and 10% levels, respectvely. postve correlaton between the certfcaton decson and frm sze that s descrbed by the logarthm of the number of employees. 6 Ths result may be plausble because larger frms usually enjoy economes of scale n

9 388 What Determnes Chnese Frms Decson on Implementng Voluntary Envronmental Schemes? mplementng certfcaton. Fnally, sector effect on the decson of eco-label certfcaton s presented n the results. We fnd sgnfcant effect n sx sectors (.e., food and beverage, metal processng and products, automoble and related products, nonmetal products, petroleum processng and cokng, and rubber and plastc products), and no sgnfcant effect n other sectors ISO Certfcaton 4 In comparson to our explanaton, another vew explaned by [11] s that foregn owners may be less wllng to contrbute to the socal well-beng of the country n whch the faclty s located and, as a result less nclned to nvest n envronmental protecton above the level of requred regulaton. However, our results do not support ths pont of vew. Ths may occur because n comparson to the case n Japan, a large share of the products produced by most foregn-owned and jont-venture frms n Chna s drectly exported abroad. Consequently, the characterstcs of ther products ncludng the pro-envronmental factor should meet what foregn consumers requre. 5 Whether a frm s lsted on a stock exchange s used to proxy shareholder pressure here. Accordng to [1], the pressures that shareholders exert over a frm arse as a result of dscontent wth envronmental fnes whch lower profts, and dsllusonment wth progress toward envronmental goals and wth dffcultes n rasng new captal or attractng new nvestors due to poor envronmental performance. Therefore, they postulated that frms that are lsted on a stock exchange are more lkely to feel such pressures and, as a result, be more lkely to ntensfy ther envronmental ntatves n order to gan favor wth or mantan ther relatons wth shareholders. However, our results do not support ther postulaton n both eco-label and ISO certfcaton decsons (see the varable of Lsted_enterprse n Tables 3 and 4). In our vew, a possble explanaton s that relatve to the envronmental performances of the lsted frms, most of Chnese nvestors n the stock market pay much more attenton to the revenues and profts of the frms. Therefore, as a result, unlke ther foregn counterparts, the lsted frms n Chna may not feel so much pressure of envronmental ssues from the shareholders. 6 We dropped the varable related to a frm s annual sales amount that could also be consdered to proxy frm sze n our fnal estmaton. Ths manpulaton s based on two consderatons. Frst, the hgh correlaton between a frm s annual sales amount and the number of employees (0.623) may cause a possble multcollnearty problem. Second, ths varable s not sgnfcant n both cases of eco-label and ISO even we dropped the varable of number of employees n the model. Table 4 presents several smlar results as those n Table 3. Among the sx desgnated reasons of certfyng ISO management standard, four of them (.e., Cost_ reducton, Envronment_load, Credblty_ncreasng, and Prestge_upgradng) exhbt strong correlaton wth the certfcaton decson, whle there seem no sgnfcant effects of rasng staffs envronmental conscousness and promotng communcatons wth outsde on the decson of ISO certfcaton. It s agan shown that foregn-owned and jont-venture enterprses are sgnfcantly more lkely to mplement ISO14001 certfcaton. Ths result s consstent wth that of [17] that foregn and jont-venture frms behave better than the Chnese frms n envronmental protecton, but dfferent from that of [11] suggestng foregn ownershp has no sgnfcant effect on Japanese frms decson of ISO14001 certfcaton. The ISO certfcaton s, as expected, sgnfcantly and postvely correlated wth the number of rvals and foregn market orentaton. In addton, same as that n the certfcaton of eco-label, there seems also to have economes of scale n the ISO certfcaton processes because of the sgnfcantly postve sgn of Frm_ sze, whch s consstent wth the results n [5] and [11] that larger frms are more lkely to undertake voluntary envronmental actons. Several dfferent results between the eco-label and ISO certfcatons are found n the followng determnants. Frst, the characterstcs of frms clent are mportant n the case of ISO Frms targetng at ndvdual consumers and/or government are more lkely to certfy ISO In addton, Other_enterprses becomes sgnfcant at 10% level f sector dummes are controlled n the model (see the last column n Table 4). Furthermore, frms belongng to the sectors of petroleum processng and cokng or rubber and plastc products are more wllng to certfy ISO 14001, whch s the same as those n the eco-label case. However, four sgnfcant sector dummes (.e., Food & beverage, Metal, Automoble, and Nonmetal) are no longer sgnfcant n the ISO case. Instead, frms n the sectors of chemcal, electrc power, and medcal products are more lkely to undertake the ISO certfcaton Margnal Effects The estmated margnal effects of each determnant of the eco-label and ISO certfcatons are summarzed n Table 5. From the table, we can see that among the reasons explanng both certfcatons, Accepted_ consumer has the hghest margnal effect on the probablty of choosng to certfy eco-label, whle Envronment_load and Cost_reducton are the two largest effects n the ISO certfcaton. It may be plausble for the former because eco-label s usually regarded as the symbol of presentng the product s pro-envronment characterstcs to consumers. In other words, whether or not the product wth the label can be accepted by consumers s a crucal ncentve for a frm choosng to seek eco-label certfcaton. In contrast, snce ISO s an nternatonal voluntary standard for envronmental management systems, frms decson to undertake the certfcaton of t normally depends on how strong ther ncentves n rasng the level of the envronmental management are. Consequently, t s not surprsng that tryng to mtgate the envronmental load caused by the frms and reducng ther producton cost become the largest two factors nfluencng the ISO certfcaton decson. Some determnants affect more the ISO certfcaton decson than those n the eco-label case. For example, the effect of foregn-owned frm on the ISO certfcaton decson (.e., 0.696) s approxmately

10 What Determnes Chnese Frms Decson on Implementng Voluntary Envronmental Schemes? tmes larger than that on the eco-label certfcaton (.e., 0.062). The smlar result s also found n the factor of foregn market orentaton. The margnal effect of the varable foregn_market on the ISO certfcaton s 0.165, whch s exactly 10 tmes larger than n the eco-label case. These fndngs may be explaned by the fact that ISO s more of an nternatonal standard, whle the eco-label s a Chnese certfcate. So frms wth nternatonal elements tend to adopt more of the ISO certfcaton. In Addton, there s no bg dfference between the effects of jont-venture frm on the choce probabltes of both certfcatons (0.219 n the eco-label case vs n the ISO case). Although the factor of frms man clents s not sgnfcant n the eco-label case, ndvdual consumer and government orentatons affect smlarly on the probablty of undertakng the ISO certfcaton, both of whch are larger than that of Other_enterprses. Furthermore, the effects of the number of rvals and frm sze on choce probablty of the ISO certfcaton are, respectvely, about 3 tmes and 4 tmes as those n the eco-label case. Wth regards to the effect of those sgnfcant sector dummes on the certfcaton decson, we fnd a bg dfference n the eco-label case. The ratos of the probablty of certfyng eco-label by frms n petroleum processng and cokng sector to those frms n other sectors are approxmately 4 for Petroleum vs Nonmetal and Automoble, 7 for Petroleum vs Food & beverage, and 17 for Petroleum vs Metal, respectvely. In comparson to these large ratos, the dfferences n the ISO case are almost neglgble. 5. Conclusons Emprcal nvestgaton on the determnants of mplementng voluntary envronmental schemes at the plant level s scarce n Chna. To shade more lght on ths subject, we conducted a frm-level survey and apply the data to study the factors nfluencng Chnese frms decson to conduct the ISO and/or Chna Envronmental Label certfcatons. Our results exhbt several smlar evdences between the factors affectng the decsons of these two certfcatons. Frst, n comparson to ther Chnese counterparts, foregn-owned and jont-venture frms are more lkely to use the certfcatons. Second, the number of rvals a frm has s postvely correlated to ts decson on the certfcatons. Thrd, foregn market orented frms are more wllng to mplement the certfcatons. Fourth, frm sze defned by the number of employees a frm has plays a sgnfcant role n determnng ts choce of dfferent certfcatons. Ffth, frms n the sectors of petroleum processng and cokng or rubber and plastc products are more lkely to certfy both ISO standard and eco-label. Sxth, whether or not a frm s lsted on a stock exchange s not a sgnfcant factor n both cases. Besdes the above mentoned smlartes, t s mportant to emphasze that we also fnd a number of dfferences n the determnants of these two certfcatons. Frst, the factor of a frm s man clent sgnfcantly affects the decson of the ISO certfcaton, but not n the eco-label case. Second, sector effect on a frm s certfcaton decson s dfferent n the certfcatons of ISO and eco-label. Fnally, except for the sector effect, other determnants nfluencng the ISO certfcaton have hgher margnal effects than those correspondng factors affectng the eco-label certfcaton. Despte the potental problem such that our sample sze may not be large enough to ensure robust estmates, our results presented n ths paper have a number of nterestng polcy mplcatons for Chnese decson makers. Frst, government polcy makers can beneft from the fact that the decsons of frms choosng to mplement voluntary envronmental schemes are nfluenced by the above specfed reasons other than regulaton. Consequently, government can provde ncentves based on these reasons to promote voluntary acton and polluton preventon as an alternatve to mandatory envronmental regulatons. By offerng approprate ncentves, programs such as ISO and Chna Envronmental Label can brng wn-wn benefts to frms, consumers, and government. Second, recognzng that varous characterstcs of frms affect ther decson makngs n undertakng voluntary envronmental programs can help government desgn more flexble polces than tradtonal regulatons such as tax and command-and-control mandates. Thrd, acknowledgng that frms n dfferent sectors have dfferent underlng motvatons n certfyng ISO and/or eco-label, Chnese government may form specfed envronmental polces based on sector specfcatons, whch as a result could be more effectve than the current undfferentated polces n Chna. Fnally, the present study s suggestve of two areas for further research. Frst, t should be noted that the results and fndngs are based on frms mostly located n the Yangtze Delta Area. Therefore, future studes may nclude frms n other areas of Chna, whch n turn could help to ensure the valdty of these fndngs. Second, varous fnancal data other than annual sales amount, whch may also be mportant n determnng frms decson of mplementng voluntary envronmental schemes, are not ncluded n the present study due to the fact that most surveyed frms refused to reveal the nformaton. We leave ths ssue as an open challenge and expect a greater effort to overcome t n follow up studes.

11 390 What Determnes Chnese Frms Decson on Implementng Voluntary Envronmental Schemes? REFERENCES [1] I. Henrques and P. Sadorsky, Envronmental Management Systems and Practces: An Internatonal Perspectve, In: N. Johnstone, Ed., Envronmental Polcy and Corporate Behavor, Edward Elgar Publshng, Northampton, USA, 2007, pp [2] ISO, The Integrated Use of Management System Standards, Internatonal Standards Organzaton, [3] J. Shen and T. Sajo, Does Energy Effcency Label Alter Consumers Purchase Decson? A Latent Class Approach on Shangha Data, Journal of Envronmental Management, Vol. 90, No. 11, 2009, pp do: /j.jenvman [4] R. Starkey, Standardzaton of Envronmental Management Systems: ISO 14001, ISO and EMAS, In: R. Welford, Ed., Corporate Envronmental Management 1: Systems and Strateges, 2nd Edton, Earthscan, London, 1998, pp [5] S. Arora and T. Cason, An Experment n Voluntary Envronmental Regulaton: Partcpaton n EPA s 33/50 Program, Journal of Envronmental Economcs and Management, Vol. 28, No. 3, 1995, pp do: /jeem [6] W. B. Gray and M. E. Dely, Complance and Enforcement: Ar Polluton Regulaton n the US Steel Industry, Journal of Envronmental Economcs and Management, Vol. 31, No. 1, 1996, pp do: /jeem [7] E. Helland, The Enforcement of Polluton Control Laws: Inspectons, Volatons, and Self-Reportng, The Revew of Economcs and Statstcs, Vol. 80, No. 1, 1998, pp do: / [8] S. Dasgupta, B. Laplante, N. Mamng and H. Wang, Inspectons, Polluton Prces, and Envronmental Performance: Evdence from Chna, Ecologcal Economcs, Vol. 36, No. 3, 2001, pp do: /s (00) [9] L. Gangadharan, Envronmental Complance by Frms n the Manufacturng Sector n Mexco, Ecologcal Economcs, Vol. 59, No. 4, 2006, pp do: /j.ecolecon [10] M. Khanna, Non-Mandatory Approaches to Envronmental Protecton, Journal of Economc Surveys, Vol. 15, No. 3, 2001, pp do: / [11] M. Nakamura, T. Takahash and I. Vertnsky, Why Japa- nese Frms Choose to Certfy: A Study of Manageral Responses to Envronmental Issues, Journal of Envronmental Economcs and Management, Vol. 42, No. 1, 2001, pp do: /jeem [12] S. B. Brunnermeer and M. A. Cohen, Determnants of Envronmental Innovaton n US Manufacturng Industres, Journal of Envronmental Economcs and Management, Vol. 45, No. 2, 2003, pp do: /s (02)00058-x [13] H. Wang, N. Mamng, B. Laplante and S. Dasgupta, Incomplete Enforcement of Polluton Regulaton: Barganng Power of Chnese Factores, Envronmental and Resource Economcs, Vol. 24, No. 3, 2003, pp do: /a: [14] S. J. DeCano and W. E. Watkns, Investment n Energy Effcency: Do the Characterstcs of Frms Matter? The Revew of Economcs and Statstcs, Vol. 80, No. 1, 1998, pp do: / [15] H. Wang and D. Wheeler, Fnancal Incentves and Endogenous Enforcement n Chna s Polluton Levy System, Journal of Envronmental Economcs and Management, Vol. 49, No. 1, 2005, pp do: /j.jeem [16] K. Fsher-Vanden, G. H. Jefferson, J. Ma and J. Xu, Technology Development and Energy Productvty n Chna, Energy Economcs, Vol. 28, No. 5-6, 2006, pp do: /j.eneco [17] H. Wang and Y. Jn, Industral Ownershp and Envronmental Performance: Evdence from Chna, Envronmental and Resource Economcs, Vol. 36, No. 3, 2007, pp [18] N. Mamng, S. Dasgupta, B. Laplante and J. H. Hong, Understandng Frms Envronmental Performance: Does News Matter? Envronmental Economcs and Polcy Studes, Vol. 9, No. 2, 2008, pp [19] S. Gupta and B. Goldar, Do Stock Markets Penalze Envronment-Unfrendly Behavour? Evdence from Inda, Ecologcal Economcs, Vol. 52, No. 1, 2005, pp do: /j.ecolecon [20] K. Elsayed and D. Paton, The Impact of Envronmental Performance on Frm Performance: Statc and Dynamc Panel Data Evdence, Structural Change and Economc Dynamcs, Vol. 16, No. 3, 2005, pp do: /j.strueco [21] Natonal Bureau of Statstcs of Chna, Chna Statstcal Yearbook, Bejng, 2007.