Consumer stated purchasing preferences and corporate social responsibility in the wood products industry: A conjoint analysis in the U.S. and China

Ecological Economics 95 (2013) 118–127

Contents lists available at ScienceDirect

Ecological Economics journal homepage: www.elsevier.com/locate/ecolecon

Surveys

Consumer stated purchasing preferences and corporate social responsibility in the wood products industry: A conjoint analysis in the U.S. and China Zhen Cai ⁎, Francisco X. Aguilar Department of Forestry, The School of Natural Resources, University of Missouri, Columbia, MO 65211, USA

a r t i c l e

i n f o

Article history: Received 28 September 2012 Received in revised form 9 August 2013 Accepted 23 August 2013 Available online 19 September 2013 Keywords: Corporate social responsibility Consumer purchasing preferences Wood products industry Conjoint analysis Hierarchical Bayesian models U.S. China

a b s t r a c t The impacts of disclosed level of corporate social responsibility (CSR), domestic versus imported origin and type of construction on consumers' stated wood product purchasing preferences were examined in the U.S. and China. Hierarchical logit models based on a Bayesian framework were utilized to test the magnitude and statistical significance of each wood product attribute using survey data. Results indicate that U.S. and Chinese respondents: (a) were more likely to choose products from manufacturing companies with a higher level of CSR rating compared with an unknown one; (b) preferred domestically manufactured wood products compared to imported ones; and (c) expressed higher interest in wood products made of solid wood compared with composites. In terms of demographics, respondents' higher education levels corresponded with higher preferences for products from companies with the highest (five-star) CSR rating in the U.S. Statistically-significant income effects were detected only in the Chinese sample when respondents indicated their purchasing preferences for wood products with three-star or five-star CSR levels. Implications for improving wood products companies' managerial performance and suggestions for future studies are provided. Published by Elsevier B.V.

1. Introduction Corporate social responsibility (CSR) conceptualizes the responsibilities businesses should maintain with society. The definition of CSR developed by the International Organization for Standardization Strategic Advisory Group on Social Responsibility (2002, p.1) describes CSR as “a balanced approach for organizations to address economic, social and environmental issues in a way that aims to benefit people, communities and society.” This definition embraces the economic, social and environmental dimensions of CSR. Economic issues under CSR encompass how companies should continuously improve their economic performance (Portney, 2005). In terms of social issues, Maignan and Ralston (2002) and Snider et al. (2003) pointed out the importance of improving local people's welfare (e.g. education level), providing good opportunities for employees for skill development and career enhancement, the provision of high-quality products and services for their customers, among others. Environmental issues covered by the CSR umbrella include: protecting water, air and soil resources and conserving biological diversity, applying energy-efficient equipment in the production process (Panwar and Hansen, 2008; Welford, 2003). In recent years, adoption of formal codes-of-conduct for CSR in the wood products industry has been discussed at length (e.g. Panwar and Hansen, 2008; Raditya, 2009; Vidal and Kozak, 2009). Compared to ⁎ Corresponding author. Tel.: +1 573 882 4295; fax: +1 573 882 1977. E-mail address: [email protected] (Z. Cai). 0921-8009/$ – see front matter. Published by Elsevier B.V. http://dx.doi.org/10.1016/j.ecolecon.2013.08.017

other industries, the wood products sector has been more heavily scrutinized for its environmental and social impacts, hence, triggering the development and implementation of formal codes for CSR (Panwar and Hansen, 2008). Among other reasons, Michael and Wiedenbeck (2004) highlighted the social and environmental importance of incorporating CSR principles along the wood products industry manufacturing chain since forests contribute to the maintenance of ecosystem health and vitality, and wood product manufacturing is often deemed to be a high-risk activity. The U.S. and China are two major wood products manufacturers where CSR has developed and been implemented differently. CSR was first embraced by the U.S. wood products industry in the 1970s with particular attention paid to environmental issues (Raditya, 2009). In recent years though, the focus of CSR seems to have shifted to social aspects of forest management and product manufacturing such as impacts on local communities and enhancement of employees' working conditions (Raditya, 2009). In China, CSR in the wood products industry was initially introduced in the 1990s (Hong and Yang, 2011) and its wide adoption has only started recently. Deforestation and environmental pollution caused along the value-added chain have been two major issues faced by the Chinese wood products industry. In response to forest conservation concerns, six major afforestation projects have been implemented since 1998 by the Chinese Central government. According to Wang and Juslin (2011), at least 7000 pulp and paper mills have closed since 1997 by China's State Environmental Protection Administration because of failure to comply with pollution regulations.

Z. Cai, F.X. Aguilar / Ecological Economics 95 (2013) 118–127

Since then, and partly as a reaction to government actions, Chinese wood products companies have proactively addressed and improved social and environmental issues associated with wood product manufacturing and embraced CSR. Past studies have examined the relationship between CSR and consumer purchasing preferences (e.g. Creyer and Ross, 1997; Pivato et al., 2008; Sen and Bhattacharya, 2001). However, there have been few academic endeavors assessing the impacts of different levels of CSR on consumer preferences. Furthermore, to our knowledge, no existing literature in the field of wood product economics has examined how different levels of CSR may influence consumer purchasing preferences. This study aims to fill this gap by exploring wood product consumers' stated reactions to CSR in the wood products industry. Specifically, we explored the U.S. and Chinese consumers' stated purchasing preferences toward wood products associated with different CSR levels, prices, type of wood product construction and domestic or imported origin. This paper is organized as follows. First, research hypotheses are developed based on the study objectives in Section 2. Section 3 describes surveys conducted in the U.S. and China, and explains data collection methods and hierarchical Bayesian (HB) models used for analysis. Section 4 presents survey results and the analyses using two HB models: one for the U.S. and the other one for China. Section 5 discusses and compares our findings with previous studies. Section 6 concludes this paper with suggestions for future research and implications for the management of wood products companies. 2. Hypotheses Development This study had three specific research objectives which were to: (a) determine the impacts of wood products companies' claimed CSR levels on stated consumer purchasing preferences; (b) parameterize consumer's stated purchasing preferences toward domestically manufactured wood products compared with imported ones; and (c) evaluate consumers' stated purchasing preferences toward wood product construction (i.e. composite and solid wood materials). These objectives guided the development of corresponding hypotheses. 2.1. CSR and Purchasing Preferences There are numerous empirical studies discussing the role that CSR initiatives play in improving consumers' views toward a company. Sen and Bhattacharya (2001) found that consumers' perceptions of CSR initiatives can influence their purchasing decisions. Mohr and Webb (2005) interviewed U.S. consumers regarding their purchasing intent for athletic shoes, finding that companies that embraced environmentallyfriendly and philanthropic manufacturing responsibilities were more likely to attract consumers' preferences compared with companies that did not. Brown and Dacin (1997) reported a positive relationship between the level of corporate giving, community involvement and respondents' evaluations toward a company's products. Creyer and Ross (1997) found that parents of elementary school children would be willing to pay price premiums for products from an ethical company. Pivato et al. (2008) interviewed organic product consumers and found that companies' CSR initiatives positively influenced consumer trust which could potentially lead to higher consumer purchasing preferences. General consumer attitudes toward products and the role of companies as environmental stewards and as members of society may differ between countries. The U.S. and China have followed different forms of economic and cultural development (Ralston et al., 1993; Shafer et al., 2007). The economy in the U.S. has developed, primarily relying on a free market. In contrast, China is experiencing an economic transition, in which state-owned enterprises remain to be dominant in the nation's economic structure, while the private sector (i.e. individual and foreign-owned enterprises) is still of a secondary order (Shafer

119

et al., 2007; Sternquist and Zhou, 1995; Suliman, 1998). In terms of culture, the historical development of the U.S. culture is based on JudeoChristian tradition with the strong premise that humans should dominate nature (O'Briant, 1974) as part of a divine mandate and right best exemplified in “Manifest Destiny” (Pratt, 1927). Confucianism, Buddhism and Taoism are three core religions or philosophies that have historically influenced the Chinese culture (Dollinger, 1988). The Chinese culture emphasizes that man is influenced and learns from nature, and should try to live harmoniously with it (Chen and Wu, 2009). Consumer attitudes toward CSR between nations may also differ because of the length of time formal codes have been in place. Given that the introduction of CSR to the Chinese wood products industry dates to about 15 years ago compared with around 40 years in the U.S., we expect wood product consumers' attitudes and purchasing preferences associated with CSR to be different. Several studies have examined how consumers are reacting to certified wood products over time. Cai and Aguilar (2013) found consumers' willingness-to-pay (WTP) price premiums for certified wood products to have increased from 1995 to 2009 after reviewing 19 consumer studies using meta-analysis. However, Ozanne and Vlosky (2003) found a decreasing consumers' WTP price premium for five wood products (a dining room set, a kitchen remodeling job, wood in a new home, stud, ready-to-assemble chair) between 1995 and 2000 in the U.S. Based on these aforementioned arguments, we hypothesized that: Hypothesis 1. Both the U.S. and Chinese consumers are more willing to purchase wood products manufactured by companies certified for being socially responsible than from companies of an unknown level of social responsibility. 2.2. Domestic Origin and Purchasing Preferences Information disclosing the origin of products can significantly influence consumer purchasing preferences (Verlegh and Steenkamp, 1999; Verlegh et al., 2005). Several studies have reported that consumers have a higher preference for domestic or locally-produced goods (Aguilar et al., 2010; Schnettler et al., 2008). Other studies have found that country of origin effects may be different between countries, often linked to their stage of economic development. For instance, Dmitrovic et al. (2009) reported that in emerging economies, consumers are on average more likely to purchase domestic products than in more advanced economies. Specific to the wood products industry, studies have found that consumer purchasing preferences can be influenced by the product region of origin. Aguilar and Vlosky (2007) found that U.S. consumers are more likely to pay a higher price for certified tropical wood products. Aguilar and Cai (2010) compared U.S. and U.K. consumers' stated wood product purchasing preferences and found that in both countries, consumers were more likely to purchase temperate wood products compared with products from tropical forests. They also concluded that U.K. consumers showed greater interests in buying temperate wood products compared with U.S. consumers. In order to explore consumers' stated purchasing preferences for wood products in terms of their country of origin, this study hypothesized that: Hypothesis 2. Both U.S. and Chinese consumers' stated preferences favor the purchase of wood products manufactured domestically. 2.3. Wood Product Construction and Purchasing Preferences The impact of type of wood product construction on consumer purchasing preferences has been discussed by Ridoutt et al. (2002), and Bowe and Bumgardner (2004), among others. Scholz and Decker (2007) studied consumer purchasing preferences toward wood furniture in Germany and found that type of wood construction (solid wood or veneer) significantly influenced consumer purchasing

120

Z. Cai, F.X. Aguilar / Ecological Economics 95 (2013) 118–127

preferences. Some studies have found that consumers expressed higher preferences for products made of solid wood rather than wood composites. For example, Anderson and Hansen (2004) found that undergraduate students at Oregon State University were more likely to purchase a CD rack made of solid wood rather than one made of composite wood. The positive effects of solid wood products on consumer stated purchasing preferences were also indicated by Jonsson et al. (2008). Hence, the last hypothesis tested in this study was: Hypothesis 3. Consumers in the U.S. and China are more willing to purchase wood products made of solid wood than products made of composite wood. 3. Methods 3.1. Questionnaire Description Two different surveys were developed and deployed in the U.S. and China. Both questionnaires had the same content except for particular demographic questions including ethnicity and income categories. The U.S. questionnaire was developed first, then translated into Chinese and translated back into English in order to ensure the consistency of the information. Pilot surveys were conducted in both countries to incorporate consumer feedback. The U.S. and Chinese questionnaires had three sections. The first section collected respondents' knowledge and perceptions of CSR. CSR was defined in this section as “the obligations of businessmen to pursue policies, make decisions, or to follow lines of action which are desirable in terms of the objectives and values of our society” (Bowen, 1953, p.270). Consumers' knowledge about CSR was collected by asking two questions (“Were you aware of CSR prior to this survey?” and “Are you aware of any companies that are socially responsible in the wood products industry?”). Respondents' past and planned future wood product consumption were attained by asking: “Have you ever purchased any wood product (e.g. wooden table, oak board) before?” and “Do you plan to purchase wood products within the next five years?” The second section focused on questions exploring consumer purchasing preferences for wood products from socially responsible companies. As suggested by Cummings and Taylor (1999) and Carlsson et al. (2005), a “cheap talk” script (Fig. 1) was presented in this section to reduce hypothetical bias. Under the cheap talk script, hypothetical bias was defined, why it exists, and respondents were asked to answer questions as an actual purchase trying to avoid such bias. Questions using a conjoint analysis (CA) were presented to elicit consumers' preferences toward selected wood product attributes after the cheap talk script. CA was initially developed by mathematical psychologists and statisticians Luce and Tukey (1964) and has been widely used in marketing research to elicit consumers' purchasing preferences (Green et al., 2001). A shopping scenario was created to simulate a real-life wood product purchasing decision. A dining table was selected

as the research subject due to its familiarity to consumers. The CA design contained multi-level attributes of a wooden dining table of the same dimensions (76 cm wide × 122 cm long × 76 cm tall). A label attached to the dining table presented information about four product attributes (wood construction, origin, CSR and price) corresponding to our research queries. Two types of wood material used to make the dining table were presented: solid wood and composite wood. Regarding origin, the dining table was assumed to be either produced domestically or imported from other countries. Due to the lack of a real CSR rating system in the wood products industry in both the U.S. and China, a conceptual CSR rating system was generated for this study. Dining table manufacturer's CSR levels were divided into three ratings and an unknown category. CSR scales had a highest rating represented by five-stars on a product label (★★★★★), three-stars captured a middle rating (★★★☆☆) and a single star represented the lowest rating (★☆☆☆☆). The idea of rating CSR levels was adapted from Waddock and Graves (1997) and stars were selected to represent companies' CSR levels in order to provide a scenario that was easy and simple for the respondents to distinguish. Respondents were informed that the rating associated with the company's CSR level was evaluated by a respected independent third-party organization and that wood products companies participated in this program voluntarily. Participants were also informed that the independent third-party organization evaluated the company according to its CSR practices including economic performance, participating environmental conservation programs, customer services, employees' working conditions and wages, and charitable donations. Three price levels: $200, $250 and $300 were set for the U.S. after searching the average price for a similar dining table in major online retailers (e.g. Walmart.com and Homedepot.com). In the Chinese survey, price levels for the dining table were converted to Chinese Yuan based on the exchange rate on September 31, 2011 (U.S. dollar: Chinese Yuan = 1: 6.38) (Financial Management Service, 2011) and rounded to the closest hundredth figure ( 1300, 1600 and 1900). We also confirmed Chinese prices to be in the price range for a similar dining table in two Chinese local furniture stores (Meilian Tiandi and Hongxing Meikailong). The attributes and their corresponding levels are summarized in Table 1. Type of wood product wood construction, country of origin, prices and CSR levels were combined into different profiles for the four-seat dining table. A fractional factorial orthogonal design was used to generate the minimum number of profiles needed to evaluate respondents' preferences (Kuhfeld et al., 1994). The generated profiles satisfied the following criteria: (1) the number of each level used in the profiles should be equal; and (2) the number of one level from one attribute together with other level from each other attribute should be the same (Caruso, 2009). Thirty-six different profiles and eighteen bundles (two profiles in each bundle) were created following Kuhfeld (2004). A status quo profile was included with each bundle as suggested by Adamowicz et al. (1998) and Louviere et al. (2000). The status quo product profile had the same attributes used in other profiles but their

Hypothetical bias: Hypothetical bias refers to the event when a respondent’s stated purchasing preference under hypothetical conditions is different from an actual purchase in a store. Why? People are likely to engage in purchasing behavior deemed to be socially preferable and avoid non-preferable products without considering potential costs. For example, when asked if they support environmental protection, many people would agree (e.g. be willing to donate money to protect the environment). However, the same people may be reluctant to donate when confronted with a real life opportunity. When actually asked to pay for something, we will first consider budget constraints. Please try to avoid this problem and respond to the following questions as if it was your actual purchase in a furniture store. Fig. 1. Cheap talk script used in the U.S. and China surveys.

Z. Cai, F.X. Aguilar / Ecological Economics 95 (2013) 118–127 Table 1 Attributes and levels for the study of consumer preferences for a wooden dining tablea used in the CA.

Wood construction Origin CSR levels

Price

U.S.

China

Composite wood Solid woodb Imported to the U.S. b Manufactured domestically Unknownb ★☆☆☆☆c ★★★☆☆c ★★★★★c $200 $250b $300

Composite wood Solid woodb Imported to Chinab Manufactured domestically Unknownb ★☆☆☆☆ ★★★☆☆ ★★★★★ 1300 1600b 1900

a

The wooden dining table is a four-seat dining table with a size of 76 cm wide × 122 cm long × 76 cm tall. b Correspond to levels used in the status quo profile in CA. c CSR ratings. ★☆☆☆☆: One-star (lowest CSR rating); ★★★☆☆: Three-star (middle CSR rating); ★★★★★: Five-star (highest CSR rating).

levels were fixed, which corresponded to solid wood, domestic product, unknown CSR and priced at $250 ( 1600). It is worth mentioning that a decision was made whether to include a status quo or an opt-out option in the CA. Empirically, we chose to include a status quo option because an opt-out alternative could have resulted in considerable loss of information. Malhotra (2008) elaborates on the risk of including an opt-out option, particularly when using on-line surveys, as it may create an incentive for respondents to respond quickly without the need to evaluate product attributes and respective trade-offs. Breffle and Rowe (2002) argue that an opt-out option can provide an opportunity to avoid the product selection task and unrealistically increase preferences toward the opt-out alternative. Nevertheless, not including an opt-out option in the choice questions may also introduce some bias to the estimated coefficients by forcing a choice (Boyle et al., 2001). Given the shortcomings of both approaches, the status quo was preferred as we deemed that the loss of information when collecting data online would have been a more serious problem than forced responses, in particular because of data collection methods included an online-based tool. Another limitation inherent to the nature of this study was the use of stated preferences instead of actual consumer behavior (Brouwer et al., 1999; Murphy et al., 2005). Actual wood product purchasing behavior should be examined in future studies through revealed preference methods,

121

which was used in several previous studies when examining consumers' WTP price premium to certified wood products (e.g. Anderson and Hansen, 2004; Anderson et al., 2005). A simulated shopping scenario, where respondents were asked to place themselves in a situation shopping for the four-seat dining table, was introduced. Shopping scenarios, where only three dining tables were available, were presented with each table carrying a label describing the four aforementioned attributes. Each respondent was asked to answer nine conjoint questions and was required to choose one dining table from each product bundle (Fig. 2). Finally, both the U.S. and Chinese surveys gathered demographic information including respondents' education, age, gross household annual income, and gender. 3.2. Data Collection The questionnaire was administered in the U.S. and China in the fall of 2011. In the U.S. an online questionnaire was administered by a sampling and data collection provider — Survey Sampling International (SSI). SSI data, with a pool of 800,000 online panelists in the U.S. has been utilized to sample this population in numerous studies (e.g. Aguilar and Cai, 2010; Thompson et al., 2006). Approximately 20,000 participants with equal distribution to female and male respondents, who were at least 18 years old, were selected across the panelists using a standard Oracle random number generation algorithm (Aguilar and Cai, 2010). A webpage link of the questionnaire was delivered to the participants and reminders were sent every 24 h by SSI. Once approximately 1000 completed responses were collected, the webpage link was turned off and online data collection stopped. An online data collection approach was dismissed to sample Chinese wood product consumers because it may not be representative of this population. Xiang (2001) argues that online respondents in China have higher annual household income and better education levels than the general population, which may introduce a significant bias in coefficient estimates compared with average Chinese wood product consumers. Therefore, in-store face-to-face interviews were selected as the preferred data collection method. The Chinese survey was conducted in three cities (Beijing, Hefei and Shenyang) to capture diverse conditions that existed across the country. Cities were chosen to capture the differences observed in the development of wood product manufacturing and ownership structure. Wood product manufacturing per capita (WPM/Capita) was used as an index to

Fig. 2. Example for the choice questions in the conjoint analysis for a four-seat dining table.

122

Z. Cai, F.X. Aguilar / Ecological Economics 95 (2013) 118–127

describe the development of the wood products industry. WPM/Capita for all of the municipalities and provinces in China were calculated.1 Then municipalities and provinces were ranked into three tier levels. Hefei city, located in Anhui province with a WPM/Capita of 0.0769 m3/capita, was placed in the highest level; Shenyang city, located in Liaoning province (0.0445 m3/capita), was in the middle tier and Beijing city (0.0049 m3/capita) was categorized in the lowest tier based on data from the National Bureau of Statistics of China (2008). Second, we aimed to sample consumers in markets with distinct types of forest/company ownership. Shenyang is located in northeastern China, where most forests and forest enterprises are state owned (Wan and Geng, 2004). Hefei is located in eastern China — forests are mainly collectively managed but forest product companies are mostly privately owned in eastern and southern China (Ding et al., 2008). The selected cities captured the mosaic of manufacturing and ownership structure in China. The survey was conducted in different furniture stores in these cities to improve representativeness of wood product consumers. To avoid potential bias we sampled consumers by their purchasing capacities (i.e. targeting only high- or low-end shops). Furniture store information (e.g. name, product price for CA product and consumer groups) in selected cities was collected using an online search engine (www.baidu. com). Furniture stores were stratified into three categories according to the average four-seat dining table price (high, medium and low). Within each stratum, stores were picked randomly. Three furniture stores in Beijing (Meilian Tiandi, IKEA and Easy Home), four in Shenyang (Baili Jiaju, Dongmaoku, IKEA and Jiulu Furniture Market) and three in Hefei (Hongxing Meikailong, Qidu Kongjian, Hongqi Jiancai) were selected.2 Respondents were approached and asked to participate in the study. Potential respondents were informed of the objectives of the survey and their willingness to participate was determined. The survey questions were then explained to participants. A response rate for this group could not be determined because of challenges with counting the exact number of people approached to participate in the study. This method did not lend itself for an accurate estimation for participation because interviewers were divided into two groups and located in different places inside the furniture stores. Several people approached who were not willing to participate might have been asked twice by interviewers, in which case the number of non-participants could have been double-counted. We should mention that the U.S. and Chinese samples may not perfectly represent their respective populations of wood product consumers because of the relatively small samples (relative to their entire country populations) and the survey administered location (particularly in China). In this regard it is worth emphasizing that results from both the U.S. and Chinese samples fit well the preferences captured in our sample, however, our findings may not be easily generalized to infer wood product preferences for an average consumer in each country.

utility that an individual i attains from choosing a dining table j from a choice set C (three alternatives) can be expressed as (Manski, 1977): U ij ¼ V ij þ εij V ij ¼ X ij β

ð1Þ

where Vij is the indirect utility function for respondent i's choice of j. Xij is a vector of attributes (price, CSR levels, wood construction and origin) for the wood product j, β is a vector of attribute parameters and εij is a random error term assumed to be independent and identically distributed. The probability (pij) that the ith individual chooses wood product j over its alternatives (any other product k (j ≠ k)) from a choice set C can be obtained indirectly as (Wilson, 2012): h i h i P ij ¼ Pr U ij NfU ik g∀k≠ j ¼ Pr V ij þ εij NfV ik þ εik g∀k≠ j

ð2Þ

  X   k ′ ′ P ij ¼ exp xij βi = 1 exp xik βi :

ð3Þ

An HB model was used to analyze the choice data as suggested by Cattin et al. (1983), Vriens et al. (1996) and Andrews et al. (2002). One of the advantages of applying HB models, over OLS or classical logit models, lies in their ability to estimate individual-level consumer data, which considers heterogeneity among individuals (i.e. each consumer may have different purchasing preferences) (Allenby and Ginter, 1995; Orme, 2000). In this study, the elicitation of aggregate-level preferences in the U.S. and China was based on the individual-level preferences, hence, HB results are an improvement over alternative econometric approaches as they control for respondent heterogeneity and avoid adding it to the model error term. The HB model had two levels: the first level corresponded to responses from all individuals in the study and the second hierarchical level referred to the behavior of a specific individual (Allenby et al., 2005). The individual i random effects {βi} in the first level were assumed to be independently and identically distributed and to follow a multivariate normal distribution with mean vector α and covariance matrix V (βi ~ Normal (α, V)) (Sawtooth, 2009). The likelihood function (Eq. (4)) was used to estimate model parameters βi in Eq. (3) following Solgaard and Hansen (2003): N

Lðfβi g; α; V Þ ¼ probðdata=fβi g; α; V Þ ¼ ∏i¼1 probðdatajβi Þpðβi ; α; V Þ ð4Þ where N is the total number of observations. Since βi ~ Normal (α, V), Eq. (4) can be rewritten as: N

Lðfβi g; α; V Þ ¼ probðdata=fβi g; α; V Þ ¼ ∏i¼1 probðdatajβi Þpðβi ; α; V Þ N

¼ ∏i¼1 Li ðβi ÞN ðβi jα; V Þ: ð5Þ According to the Bayesian theorem, Eq. (5) could be rewritten as:  N p fβi g; α; V jdataÞ∝∏i¼1 Li ðβi ÞNðβi jα; V Þpðα; V Þ

ð6Þ

3.3. Data Analysis To better capture consumers' purchasing preferences, only respondents who indicated that they have purchased wood products in the past and were planning to purchase wood products within the next five years were included in the econometric analysis. The estimation of consumers' stated purchasing preferences for the four-seat dining table was based on a random utility model (RUM). RUM indicates the

1 Provincial level data was used as a proxy for city level data since wood product information was only provided at the province level except for the municipalities in China. All data were sourced from National Bureau of Statistics of China (2008). 2 The stratification of stores as per average prices was as follows. Low-price tier: Meilian Tiandi, Dongmaoku, Jiulu Furniture Market, Hongqi Jiancai. Medium-tier: IKEA and Qidu Kongjian. High-price tier: Baili Jiaju, Easy Home and Hongxing Meikailong.

where p({βi},α, V data) is the posterior distribution and p(α, V) is the prior distribution. α is normally distributed and has means equal to the averages of each beta. The V follows an inverted Wishart distribution (Solgaard and Hansen, 2003; Train, 2001). Software Winbugs (the MS Windows operating system version of Bayesian Analysis Using Gibbs Sampling) was used to estimate the mean for each parameter through the Gibbs sampling method (Lunn et al., 2000). To estimate parameters of the HB model, 20,000 random draws (iterations) were calculated. The first “Burn in” 10,000 iterations were not included in the estimation to eliminate initial draw effects (Spiegelhalter et al., 2003). The thinning rate for the estimated models was set as 10 in order to avoid autocorrelation within the generated samples. Convergence of the Gibbs sampling was checked using the Brooks–Gelman–Rubin convergence diagnostic (Brooks and Gelman, 1997). Monte Carlo (MC)

Z. Cai, F.X. Aguilar / Ecological Economics 95 (2013) 118–127

123

Table 2 Model dependent and independent variables and descriptions. Variable

Description

Dependent variable Respondents' choice for each profile

1 = if the profile was chosen by the respondent, 0 = otherwise

Independent variables Dining table attributes Solid Domestic Price Unknown One-star CSR level Three-star CSR level Five-star CSR level Characteristic specific variables One-star × gender One-star × agea One-star × incomeb One-star × educationc Three-star × gender Three-star × age Three-star × income Three-star × education Five-star × gender Five-star × age Five-star × income Five-star × education

1 = if the dining table is made of solid wood, 0 = otherwise 1 = if the dining table was produced domestically, 0 = otherwise Continuous variable. $200, $250, $300 or 1300, 1600, 1900. Prices used in the models were standardized. Base line variable indicating the CSR level is unknown 1 = if the CSR level was ranked One-star, 0 = otherwise 1 = if the CSR level was ranked Three-star, 0 = otherwise 1 = if the CSR level was ranked Five-star, 0 = otherwise Interaction variable, 1 = if the CSR level was ranked One-star and the respondent is male, 0 = otherwise Interaction variable between the One-star CSR level and respondents Age Interaction variable between the One-star CSR level and respondents Income Interaction variable between the One-star CSR level and respondents Education Interaction variable, 1 = if the CSR level was ranked Three-star and the respondent is male, 0 = otherwise Interaction variable between the Three-star CSR level and respondents Age Interaction variable between the Three-star CSR level and respondents Income Interaction variable between the Three-star CSR level and respondents Education Interaction variable, 1 = if the CSR level was ranked Five-star and the respondent is male, 0 = otherwise Interaction variable between the Five-star CSR level and respondents Age Interaction variable between the Five-star CSR level and respondents Income Interaction variable between the Five-star CSR level and respondents Education

a Age: 1 indicates that the age of the respondent is younger than 25; 2 indicates that the age of the respondent is between 26 and 35; 3 indicates that the age of the respondent is between 36 and 45; 4 indicates that the age of the respondent is between 46 and 55; and 5 indicates that the age of the respondent is older than 55. b Annual household income: 1 = respondent's annual household income b$24,999 (U.S.), 24,999 (China); 2 = respondent's annual household income between $25,000–49,999 (U.S.), 25,000–49,999 (China); 3 = respondent's annual household income between $50,000–74,999 (U.S.), 50,000–74,999 (China); 4 = respondent's annual household income between $75,000–99,999 (U.S.), 75,000–99,999 (China); 5 = respondent's annual household income between $100,000–124,999 (U.S.), 100,000–124,999 (China); 6 = respondent's annual household income more than $125,000 (U.S.), 125,000 (China). c Education: 1 indicates that the respondent has a high school degree; 2 indicates that the respondent has some college education; 3 indicates that the respondent has a college degree; 4 indicates that the respondent has a graduate degree; and 5 indicates others.

errors for each parameter were estimated to determine the difference between the sample means and true posterior means. Parameter odds ratios (OR) were calculated using Eq. (7): βi

ORðβi Þ ¼ exp :

ð7Þ

Two models were generated to explore consumer stated purchasing preferences in the U.S. and China. Variables used in the models are presented in Table 2. Wood product consumers' WTP price premiums for different levels of CSR compared with an unknown CSR level, domestically produced wood products compared with imported ones, and solid wood products compared with composited ones were calculated following Cai and Aguilar (2013): WTP price premiums for one−starðthree−star; five−starÞCSR level ð8Þ compared with unknown CSR level ¼ Coefficient of one−star ðthree−star; five−starÞCSR level=Price Coefficient WTP price premiums for solid wood products compared with composite ones ¼ Coefficient of solid wood products=Price Coefficient ð9Þ WTP price premiums for domestically produced wood products ð10Þ compared with importedones ¼ Coefficient of domestically produced wood products=Price Coefficient:

4. Results 4.1. Survey Datasets A total of 2012 responses were collected, 1120 responses from the U.S. and 892 from China (328 observations from Beijing, 219 from Hefei and 345 from Shenyang). Approximately 83% of the respondents in the U.S.

and 84% in China indicated that they have purchased wood products before taking the survey and nearly half (54%) of the U.S. respondents compared to approximately three-fourths (73%) of the Chinese respondents stated they will purchase wood products within the next five years. After screening our respondents based on their past and stated future planned wood product purchases, 545 respondents in the U.S. and 540 Chinese respondents were left and their responses were analyzed using the HB models. Table 3 presents general statistics that describe all collected U.S. and Chinese respondents, and screened samples. 4.2. CSR Knowledge Among U.S. consumers, 37% indicated they had heard about CSR before taking the survey compared with 60% in China. The reason for the higher rate in China might be the ubiquitous government information campaign promoting CSR through public media. Arguably, Chinese respondents' previous CSR knowledge may be reflected on a higher level of awareness about CSR in the wood products industry compared with the U.S. sample. When asked if they have heard about any companies that are socially responsible in the wood products industry, 14% of the U.S. respondents indicated their awareness compared with 26% in China. 4.3. Hierarchical Bayesian (HB) Models HB model results (including odds ratio, standard deviation and MC error) are presented in Table 4.3 The MC error for each parameter was 3 Choice data was also estimated through classical random parameter logit regression models. In terms of variable significance and variable signs, results from the Bayesian model and classic models were very similar except for few differences for some interaction variables. Several differences regarding the magnitudes of independent variables existed. Given limited space, estimated results from classical random parameter logit regression models are not presented in the paper. Results are available upon request.

124

Z. Cai, F.X. Aguilar / Ecological Economics 95 (2013) 118–127

Table 3 Distribution of the U.S. and Chinese respondents' demographic information. Screened consumersa

All respondents

Gender Female Male Age (years) b25 26–35 36–45 46–55 N55 Annual household income (Chinese category is listed in parenthesis) b$24,999 (b 23,999) $25,000–$49,999 ( 24,000– 47,999) $50,000–$74,999 ( 48,000– 71,999) $75,000–$99,999 ( 72,000– 95,999) $100,000–$124,999 ( 96,000– 119,999) N$125,000 (N 120,000) Rather not say Education High school Some college College Graduate Other a

U.S. (n = 1120)

China (n = 892)

U.S. (n = 545)

China (n = 540)

48.77% 51.23%

55.67% 44.33%

48.99% 51.01%

54.26% 45.74%

11.16% 14.55% 12.32% 18.48% 43.48%

28.48% 34.87% 19.84% 11.66% 5.16%

11.56% 20.18% 13.94% 22.57% 31.74%

23.52% 38.89% 19.81% 11.30% 6.48%

21.13%

15.47%

17.43%

13.52%

30.76%

27.69%

30.46%

26.30%

20.50%

20.52%

22.39%

21.82%

9.98%

8.63%

12.66%

9.07%

4.59%

13.68%

5.87%

13.52%

5.67% 7.37%

10.43% 3.59%

6.79% 4.40%

12.78% 2.96%

26.65% 30.89% 24.57% 17.07% 0.81%

18.67% 25.57% 42.19% 11.88% 1.70%

22.02% 30.64% 28.07% 18.53% 0.73%

17.59% 25.74% 41.30% 14.26% 1.11%

Screened dataset included respondents who indicated that they had purchased wood products in the past and were willing to purchase in the next five years.

less than about 5% of the sample standard deviation suggesting that 10,000 iterations were appropriate. Statistically significant variables are identified with asterisks (*) in Table 4. Negative parameters are indicated by including a negative sign (−) before the odds ratio for the corresponding parameter. HB model results from both the U.S. and China indicate that both the three- and five-star CSR levels were statistically significant compared with the status quo of unknown information. Consumers from both countries reported their preferences toward buying dining tables that are produced by at least a three-star CSR level company compared with an unknown CSR level company. U.S. consumers were 5.93 times more likely to purchase the same dining table manufactured by a three-star CSR level company than the dining table made by an unknown CSR level company, and were 26.71 times more likely to buy the dining table produced from a five-star CSR level company than an unknown CSR company. Chinese consumers' stated preferences toward the products from three-or fivestar CSR level company were only about 5–10 times more than from an unknown one. Results show that a dining table manufactured by one-star CSR level company had no effects on consumer statedproduct choice compared with the status quo in the U.S., while the lower CSR level had a positive effect in China. U.S. consumers were 22.38 times more likely to choose a solid-wood dining table rather than one made of composite materials, while Chinese consumers were only 6.44 times more likely to select a table made of solid wood. These findings support the hypothesis that consumers from both countries are willing to buy domestic-manufactured wood products. U.S. consumers were also 26.72 times more likely to select a dining table manufactured in-country, compared with 9.64 times in China — both compared to an imported one. Respondent-specific characteristics on stated purchasing preferences were detected. Consumer preferences toward purchasing a dining table from CSR companies exhibited education effects in the U.S. sample when choosing a wood product with a five-star CSR level. A higher level of income resulted in a higher probability for consumers to buy wood

products from three- or five-star CSR companies in China. No income effects were detected in the U.S. Welfare estimates were also conducted based on model results. WTP results indicate that U.S. consumers were willing to pay more for a dining table manufactured by companies that had CSR ratings of three($10.13) or five-stars ($17.37) compared with an unknown CSR level. Chinese consumers' WTP price premiums were $14.17 and $17.73 for a dining table with a three- and five-star CSR rating, respectively. The U.S. consumers were willing to pay around $16.70 more for a wooden dining table made by solid wood compared with composite wood, or a dining table manufactured domestically compared with an imported one. However, Chinese participants were only willing to pay $3.30 more for a wooden dining table manufactured domestically. Compared with a dining table made of composite wood, Chinese consumers were willing to pay around $15.04 more for a solid dining table. When comparing WTP associated with all product attributes, a five-star CSR level captured the highest premium, predominantly in the Chinese sample. However, the results associated with comparisons between countries should be interpreted with caution since there might be some estimation biases caused by different sampling and survey administration methods. 5. Discussion In both the U.S. and China, respondents reported preferences toward buying wood products manufactured from three-or five-star CSR rated companies rather than those of unknown CSR level. The results are consistent with Brown and Dacin (1997), and Creyer and Ross (1997), who reported that companies' CSR initiatives have a positive relationship with consumer purchasing preferences. Specific to the wood products industry, our results are also consistent with Anderson and Hansen (2004) and Aguilar and Cai (2010) who found that consumers are more likely to choose certified wood products than non-certified ones. U.S. and Chinese respondents' stated preferences for wood products with an unknown CSR level compared with one-star CSR level companies

Z. Cai, F.X. Aguilar / Ecological Economics 95 (2013) 118–127

125

Table 4 Odds ratio, standard deviation and MC error of hierarchical Bayesian models.

Dining table attributes Solid Domestic Price One-star × CSR level Three-star × CSR level Five-star × CSR level Characteristic specific variables One-star × gender One-star × age One-star × income One-star × education Three-star × gender Three-star × age Three-star × income Three-star × education Five-star × gender Five-star × age Five-star × income Five-star × education a

Model 1

Model 2

U.S. (n = 545)

China (n = 540)

Odds ratio

Standard deviation

MC error

Odds ratio

Standard deviation

MC error

24.4102a 23.3828a (−)0.0017a 1.0899 6.9379a 27.7157a

0.1686 0.1610 0.4102 0.3484 0.3222 0.2045

0.0139 0.0130 0.0378 0.0347 0.0319 0.0199

7.4409a 1.5530a (−)0.0117a 2.3686a 6.6259a 10.6440a

0.1245 0.1148 0.3005 0.2953 0.3071 0.3008

0.0089 0.0091 0.0259 0.0293 0.0305 0.0296

1.3266 1.0482 (−)0.7952 1.1449 1.0776 1.0835 (−)0.9933 1.2728 (−)0.9122 1.1747 0.9306 1.4560a

0.1882 0.0895 0.1116 0.1340 0.1737 0.0856 0.0891 0.1181 0.2184 0.1157 0.1098 0.1465

0.0183 0.0083 0.0104 0.0128 0.0169 0.0080 0.0084 0.0113 0.0214 0.0110 0.0104 0.0140

(−)0.8994 1.1499 1.0881 1.0008 (−)0.8252 1.0581 1.3053a 1.2278 (−)0.9399 1.1003 1.7018a 1.1737

0.1903 0.1130 0.0688 0.1236 0.2124 0.0907 0.0715 0.1080 0.1926 0.1166 0.1166 0.1257

0.0186 0.0108 0.0061 0.0118 0.0208 0.0084 0.0066 0.0103 0.0187 0.0110 0.0111 0.0119

Indicated that the parameter of the corresponding variable significantly influences consumers' purchasing preferences.

were different. The explanation for the difference may be elucidated by examining the context in which consumers purchase their wood products in each nation. We argue that Chinese respondents may have perceived that a product from a one-star rated company certified by an independent organization should at least meet minimum national wood product manufacturing regulations and laws in order to participate in the CSR program. For wood products companies that have no proof of CSR participation, their manufacturing practices are unknown and Chinese consumers may show some level of concern about their behavior and potential lack of consistency with existing regulations. In the U.S., respondents may not have this concern. We argue that U.S. respondents expect companies with an unknown CSR level to meet applicable laws and regulations and have the same responsibilities as single-star CSR rated companies. Our findings also suggest that both U.S. and Chinese sampled consumers reported strong preferences for domestically manufactured wood products. The U.S. result is consistent with the findings of Donovan and Nicholls (2003) who found that respondents are willing to pay an additional $82 for an end table made in Alaska compared with the same one made in China, holding everything else constant. Our results are also congruent with findings reported by Verlegh and Steenkamp (1999), Verlegh et al. (2005) and Schnettler et al. (2008). Real market data in the U.S. may shed some light on the practical implications of this result. Most recently available market data show that approximately 29% of the total U.S. wood and paper product consumption was imported (USDA Forest Service, 2010) and about 40% of all wood residential furniture sold in the U.S. was imported — with projected increases (Buehlmann et al., 2003). Buehlmann et al. (2003) identified the lower price of wood product imports as their main competitive advantage. Our results suggest that as imported wood products lose their price competitive advantage due to rising labor and transportation cost, the U.S. consumers would exhibit higher preferences for domestically produced wood products in the future. In the meantime, U.S. manufacturers could partly offset higher prices with disclosure of domestic origin and CSR compliance. Market data on wooden furniture production and imports in China shows that from January to July 2009 imported wooden furniture accounted for approximately 1.44% of China's total production, and only 9.98% of the wooden furniture consumed in China was imported (China Commodity Market Place, 2009). Such a small share of imported furniture in China may point to the lack of familiarity of local consumers with foreign-made wood products

and the comparatively high prices for imported wood products. We foresee that as the share of imports might increase in the future, the magnitude of the domestic origin coefficient may change as well. We suggest two reasons to help explain a consistent higher level of preference toward domestic wood products in both countries. First, it might be associated with consumer ethnocentrism (Balabanis and Diamantopoulos, 2004; Evanschitzky et al., 2008). Consumer ethnocentrism describes how attitudes favoring purchasing local products might be based on moral grounds (Verlegh and Steenkamp, 1999). Consumers, who are more ethnocentric, are more likely to worry about the perceived negative social impact that imported products have on the home country, such as unemployment, and favor buying domestic products. Second, greater preferences for in-country manufactured products may be linked to consumers' stereotyped country image (Lotz and Hu, 2001). Consumers' country image is usually formed by the country's “representative products, national characteristics, economic and political background, history and traditions” (Nagashima, 1970, p.68). Consumers judge products from different countries or regions to be of greater or lesser appeal compared to domestic ones based on a country's image. Though, this image may not always be positive and it may change from one country to another. In the case of our study, no specific country of origin was identified for imported products' limiting consumers' comparative product evaluation based on two countries' perceptions. Our findings are only applicable to a U.S.-or China-manufactured dining table against a general imported one. Coefficients nonetheless are expected to change in magnitude and even sign if another specific origin of imported country is identified. Higher consumer preferences for purchasing a domestic wooden dining table may provide some practical implications for company managers and policy makers. The U.S. is the world's leading wood products importer and its domestic wood product manufacturing sector is facing major competition from abroad (Woodall et al., 2012). Wood products imported to the U.S. are often relatively cheaper due to current lower raw material and labor costs in developing economies (primarily in Asia). The U.S. domestic wood product manufacturers may strategically emphasize their products' origin and its “Made in the U.S.A.” label to enhance competing with foreign producers. Even if a product was not assembled in the U.S. but made of U.S. wood, it might also capture some benefits from the market. Consumers in China also expressed higher interests in buying a domestic manufactured wooden dining table compared against an imported one. While in the U.S., above-ground

126

Z. Cai, F.X. Aguilar / Ecological Economics 95 (2013) 118–127

timber volume (net of removals and mortality) continues to grow (USDA Forest Service, 2011a), forest resources in China are under great pressure from its large population and the increasing economy (Zhang and Gan, 2007). Chinese policy makers should take an active approach to encourage timber procurement from companies with formal CSR codes-of-conduct regardless of domestic or foreign origin. Results from the CA also suggest higher consumer interest in buying a dining table made of solid wood rather than composites, keeping other wood product attributes constant. This result is consistent with Anderson and Hansen (2004) and Jonsson et al. (2008), both of whom indicated that consumers are more likely to purchase solid wood products. Consumers' preferences for solid wood products are reflected on market data. In 2009, the U.S. wood product consumption comprised of softwood and hardwood lumber totaled 75.11 million cubic meters, while composite wood product consumption amounted to 35.09 million cubic meters (USDA Forest Service, 2011b).

6. Conclusions This study used a CA to determine U.S. and Chinese consumers' stated purchasing preferences toward different wood product attributes. Specifically, we explored the impact of wood construction, domestic versus imported origin, price and wood products companies' CSR levels on consumer purchasing preferences for a dining table. An HB logit model was utilized to estimate the sign, magnitude and MC errors of product-specific attributes, and participants' demographics. Model results from samples of U.S. and Chinese consumers indicate that higher CSR ratings increased stated purchasing preferences. A dining table manufactured by a three- or five-star rated CSR company was much more preferred than one from a company of unknown CSR level. Consumers from both countries indicated higher levels of preference for solid wood dining table compared with one of composite construction. Both U.S. and Chinese consumers preferred a domestic dining table compared with an imported one. Respondents' education levels influenced their stated choices with regard to higher CSR level companies in both countries. In conclusion, our data suggests that higher CSR rating resulted in higher stated consumer preferences. Overall there was a higher likelihood for selecting a dining table made of solid wood compared to one made of composites; and there were significant price effects on stated product choices in both the U.S. and China samples. Between country samples, our results indicate that consumers in the U.S. and China viewed a dining table from companies with a one-star CSR rating level differently. Based on our results, wood products companies in both the U.S. and China should be motivated to disclose all this information when products exhibit favorable attributes (e.g. high CSR rating, solid wood construction, domestically manufactured) to expand market shares. Nevertheless, because our results were generated using different survey methods, future studies should further test country differences by applying the same data collection methods in the two countries. Particularly, further research should concentrate on the examination of the effects of different CSR levels across countries and how domestic product preferences change when labels identify the country exporting wood products to the U.S. or China.

Acknowledgments This study was partly made possible by a John Bies International Travel Scholarship awarded by the University of Missouri Graduate School. The authors thank Dr. Hong He from the University of Missouri, Mr. Yu Wan from China Unicom and Ms. Liang Zhang from ZhongTai ShengRui Ltd. China for their support to collect data in China. We also acknowledge editorial support from Dr. Bill Kurtz, Emeritus Professor at the University of Missouri.

References Adamowicz, W., Louviere, J., Swait, J., 1998. Introduction to attribute-based stated choice methods. Final report to Resource Valuation Branch, Damage Assessment Center, National Oceanic and Atmospheric Administration, U.S. Department of Commerce, Advanis, Edmonton, Canada (44 pp.). Aguilar, F.X., Cai, Z., 2010. Conjoint effect of environmental labeling, disclosure of forest of origin and price on consumer preferences for wood products in the US and UK. Ecol. Econ. 70 (2), 308–316. Aguilar, F.X., Vlosky, R.P., 2007. Consumer willingness to pay price premiums for environmentally certified wood products in the U.S. For. Policy Econ. 9, 1100–1112. Aguilar, F.X., Cernusca, M.M., Gold, M.A., Barbieri, C., 2010. Frequency of consumption, familiarity and preferences for chestnuts in Missouri. Agroforest. Syst. 79, 19–29. Allenby, G.M., Ginter, J.L., 1995. Using extremes to design products and segment markets. For. Policy Econ. 32 (4), 392–403. Allenby, G.M., Rossi, P.E., McCulloch, R.E., 2005. Hierarchical Bayes Models: A Practitioners Guide. Available at SSRN: http://ssrn.com/abstract=655541. http://dx.doi.org/10.2139/ ssrn.655541 (Last accessed: April 29, 2012). Anderson, R.C., Hansen, E.N., 2004. The impact of environmental certification on preferences for wood furniture: a conjoint analysis approach. For. Prod. J. 54 (3), 42–51. Anderson, R.C., Laband, D.N., Hansen, E.N., Knowles, C.D., 2005. Price premiums in the mist. For. Prod. J. 55 (6), 19–22. Andrews, R.L., Ansari, A., Currim, I.S., 2002. Hierarchical Bayes versus finite mixture conjoint analysis models: a comparison of fit, prediction, and partworth recovery. J. Mark. Res. 39 (1), 87–98. Balabanis, G., Diamantopoulos, A., 2004. Domestic country bias, country-of-origin effects, and consumer ethnocentrism: a multidimensional unfolding approach. J. Acad. Mark. Sci. 32 (1), 80–95. Bowe, S.A., Bumgardner, M.S., 2004. Species selection in secondary wood products: implications for product design and promotion. Wood Fiber Sci. 34 (3), 408–418. Bowen, H.R., 1953. Social Responsibilities of the Businessman. Harper&Row, NewYork. Boyle, K.J., Holmes, T.P., Teisl, M.F., Roe, B., 2001. A comparison of conjoint analysis response formats. Am. J. Agric. Econ. 83 (2), 441–454. Breffle, W.S., Rowe, R.D., 2002. Comparing choice question formats for evaluating natural resource tradeoffs. Land Econ. 87 (2), 298–314. Brooks, S.P., Gelman, A., 1997. General methods for monitoring convergence of iterative simulations. J. Comput. Graph. Stat. 7, 434–455. Brouwer, R., Langford, I.H., Bateman, I.J., Turner, R.K., 1999. A meta-analysis of wetland contingent valuation studies. Reg. Environ. Change 1 (1), 47–57. Brown, T.J., Dacin, P.A., 1997. The company and the product: corporate associations and consumer product responses. J. Mark. 61, 68–84. Buehlmann, U., Bumgardner, M., Schuler, A., Rich, C., 2003. In: Deisenroth, Michael P. (Ed.), Enterprise Integration, a Production Perspective, Electronic Proceedings of the 17th International Conference on Production Research; 2003 August 3–7; Blacksburg, VA. Virginia Polytechnic Institute, Blacksburg, VA. Cai, Z., Aguilar, F.X., 2013. Meta-analysis of consumer's willingness-to-pay estimates for certified wood products. J. For. Econ. 19, 15–31. Carlsson, F., Frykblom, P., Lagerkvist, C.J., 2005. Using cheap-talk as a test of validity in choice experiments. Econ. Lett. 89, 147–152. Caruso, E.M., 2009. Using conjoint analysis to detect discrimination: revealing covert preferences from overt choices. Soc. Cogn. 27 (1), 128–137. Cattin, P., Gelfand, A.E., Danes, J., 1983. A simple Bayesian procedure for estimation in a conjoint model. J. Mark. Res. 20 (1), 29–35. Chen, X., Wu, J., 2009. Sustainable landscape architecture: implications of the Chinese philosophy of “unity of man with nature” and beyond. Landscape Ecol. 24 (8), 1015–1026. China Commodity Market Place, 2009. Pine furniture market analysis report and predictions for 2010–2013. Available online at: http://baogao.chinaccm.com/48/4810/ 481003/news/20091224/121538.asp#dinggou (Last accessed: March, 21, 2013. In Chinese). Creyer, E.H., Ross Jr., W.T., 1997. The influence of firm behavior on purchase intention: do consumers really care about business ethics. J. Consum. Mark. 14 (6), 421–432. Cummings, R.G., Taylor, L.O., 1999. Unbiased value estimates for environmental goods: a cheap talk design for the contingent valuation method. Am. Econ. Rev. 89 (3), 649–665. Ding, Q.J., Mack, R.N., Lu, P., Ren, M.X., Huang, W.H., 2008. China's booming economy is sparking and accelerating biological invasions. Bioscience 58 (4), 317–324. Dmitrovic, T., Vida, I., Reardon, J., 2009. Purchase behavior in favor of domestic products in the West Balkans. Int. Bus. Rev. 18 (5), 523–535. Dollinger, J., 1988. Confucian ethics and Japanese management practices. J. Bus. Ethics 7, 575–584. Donovan, G.H., Nicholls, D.L., 2003. Estimating consumer willingness to pay a price premium for Alaska secondary wood products. Res. Pap. PNWRP- 553. U.S. Department of Agriculture, Forest Service, Pacific Northwest Research Station, Portland, OR (7 pp.). Evanschitzky, H., Wangenheim, F., Woisetschläger, D., Blut, M., 2008. Consumer ethnocentrism in the German market. Int. Mark. Rev. 25 (1), 7–32. Financial Management Service, 2011. Treasury reporting rates of exchange. Available online at: http://www.fms.treas.gov/intn_09-30-2011.pdf (Last accessed: April. 25, 2012). Green, P.E., Krieger, A.M., Wind, Y., 2001. Thirty years of conjoint analysis: reflections and prospects. Interfaces 31, 56–73. Hong, X., Yang, X.H., 2011. Review of corporate social responsibility in China. China Bus. Trade 12, 255–256 (In Chinese). International Organization for Standardization, 2002. Strategic advisory group on corporate social responsibility: preliminary working definition of organizational social responsibility. ISO/TBM AGCSR N4. International Organization for Standardization, Geneva.

Z. Cai, F.X. Aguilar / Ecological Economics 95 (2013) 118–127 Jonsson, O., Lindberg, S., Roos, A., Hugosson, M., Lindstrom, M., 2008. Consumer perceptions and preferences on solid wood, wood-based panels, and composites: a repertory grid study. Wood Fiber Sci. 40 (4), 663–678. Kuhfeld, W.F., 2004. Marketing research methods in SAS. Available online at http:// support.sas.com/techsup/technote/ts689.pdf (Last accessed: April 21, 2012). Kuhfeld, W.F., Tobias, R.D., Garratt, M., 1994. Efficient experimental design with marketing research applications. J. Mark. Res. 31, 545–557. Lotz, S.L., Hu, M.Y., 2001. Diluting negative country of origin stereotypes: a social stereotype approach. J. Mark. Manag. 17, 105–135. Louviere, J.J., Hensher, D.A., Swatt, J.D., 2000. Stated Choice Methods: Analysis and Application. Cambridge University Press. Luce, R.D., Tukey, J.W., 1964. Simultaneous conjoint measurement. J. Math. Psychol. 1, 1–27. Lunn, D.J., Thomas, A., Best, N., Spiegelhalter, D., 2000. WinBUGS — a Bayesian modeling framework: concepts, structure, and extensibility. Stat. Comput. 10, 325–337. Maignan, I., Ralston, D.A., 2002. Corporate social responsibility in Europe and the US: insights from businesses self-presentations. J. Int. Bus. Stud. 33 (3), 497–514. Malhotra, N., 2008. Completion time and response order effects in web surveys. Public Opin. Q. 72 (5), 914–934. Manski, C., 1977. The structure of random utility models. Theory Decis. 8, 229–254. Michael, J.H., Wiedenbeck, J.K., 2004. Safety in the wood products industry. Available online at: http://www.fs.fed.us/ne/newtown_square/publications/other_publishers/ OCR/ne_2004_michael001.pdf (last accessed: May, 12, 2012). Mohr, L.A., Webb, D., 2005. The effects of corporate social responsibility and price on consumer responses. J. Consum. Aff. 39 (1), 121–147. Murphy, J.J., Allen, P.G., Stevens, T.H., Weatherhead, D., 2005. A meta-analysis of hypothetical bias in stated preference valuation. Environ. Resour. Econ. 30, 313–325. Nagashima, A., 1970. A comparison of Japanese and U.S. attitudes towards foreign products. J. Mark. 34, 68–74. National Bureau of Statistics of China, 2008. Forest Resource Distribution in China. Available online at: http://www.stats.gov.cn/tjsj/qtsj/hjtjzl/hjtjsj2010/t20111228_402788786. htm, (Last accessed: February 21, 2013. In Chinese). O'Briant, W.H., 1974. Man, nature, and the history of philosophy. In: Blackstone, W.T. (Ed.), Philosophy and Environmental Crisis. University of Georgia Press, Athens, pp. 79–89. Orme, B., 2000. Hierarchical Bayes: Why all the attention? Quirk's Marketing Research Review. March 2000. Available online at: http://www.sawtoothsoftware.com/ download/techpap/hbwhy.pdf (Last accessed: March, 21, 2013). Ozanne, L.K., Vlosky, R.P., 2003. Certification from the US consumer perspective: a comparison of 1995 and 2000. For. Prod. J. 53 (3), 13–21. Panwar, R., Hansen, E., 2008. CSR in the forest sector. UNSYLVA: An International Journal, 230(59). Food and Agriculture Organization (FAO) of the United Nations, pp. 45–48. Pivato, S., Misani, N., Tencati, A., 2008. The impact of corporate social responsibility on consumer trust: the case of organic food. Bus. Ethics 17, 3–12. Portney, P.R., 2005. Corporate social responsibility: an economic and policy perspective. In: Hay, Bruce L., Stavins, Robert N., Vietor, Richard K. (Eds.), Environmental Protection and the Social Responsibility of Firms. Resources for the Future, Washington, DC, pp. 107–131. Pratt, J.W., 1927. The origin of “Manifest Destiny”. Am. Hist. Rev. 32, 795–798. Raditya, D.A., 2009. Case studies of corporate social responsibility (CSR) in forest products companies — and customer's perspectives. Second Cycle, A1E.SLU, Dept. of Forest Products, Uppsala. Ralston, D.A., Gustafson, D.J., Cheung, F., Terpstra, R.H., 1993. Differences in managerial values: a study of U.S., Hong Kong and PRC managers. J. Int. Bus. Stud. 24 (2), 249–275. Ridoutt, B.G., Ball, R.D., Killerby, S.K., 2002. Wood in the interior office environment: effects on interpersonal perception. For. Prod. J. 52 (9), 23–30. Sawtooth Software, Inc., 2009. Software for hierarchical Bayes estimation for CBC data. Available online at: http://www.sawtoothsoftware.com (Last accessed: April 29, 2012). Schnettler, B., Ruiz, D., Sepulveda, O., Sepulveda, N., 2008. Importance of the country of origin in food consumption in a developing country. Food Qual. Preference 19 (4), 372–382.

127

Scholz, S.W., Decker, R., 2007. Measuring the impact of wood species on consumer preferences for wooden furniture by means of the Analytic Hierarchy Process. For. Prod. J. 57 (3), 23–28. Sen, S., Bhattacharya, C.B., 2001. Does doing good always lead to doing better? Consumer reactions to corporate social responsibility. J. Mark. Res. 38 (2), 225–243. Shafer, W.E., Fukukawa, K., Lee, G.M., 2007. Values and the perceived importance of ethics and social responsibility: the US versus China. J. Bus. Ethics 70 (3), 265–284. Snider, J., Hill, R.P., Martin, D., 2003. Corporate social responsibility in the 21st century: a view from the world's most successful firms. J. Bus. Ethics 48 (2), 175–187. Solgaard, H.S., Hansen, T., 2003. A hierarchical Bayes model of choice between supermarket formats. J. Retail. Consum. Serv. 10 (3), 169–180. Spiegelhalter, D., Thomas, A., Best, N., Lunn, D., 2003. WinBUGS User Manual, Version 1.4. MRC Biostatistics Unit, Cambridge, UK. Sternquist, B., Zhou, X.Q., 1995. China: the planned to face market paradigm. Int. J. Retail Distrib. Manag. 23, 21–28. Suliman, M.O., 1998. China's Transition to a Socialist Market Economy. Quorum Books, Westport, CN. Thompson, M.E., Fong, G.T., Hammond, D., Boudreau, C., Driezen, P., Hyland, A., Borland, R., Cummings, K.M., Hastings, G.B., Siahpush, M., Mackintosh, A.M., Laux, F.L., 2006. Methods of the international tobacco control (ITC) four country survey. Tob. Control. 15, 12–18. Train, K., 2001. A comparison of hierarchical Bayes and maximum simulated likelihood for mixed logit. Paper Presented at the Asilomar Invitational Choice Symposium, California, June. USDA Forest Service, 2010. Criteria and indicators for forest sustainability. Available online at: http://www.fs.fed.us/research/sustain/criteria-indicators/indicators/indicator-632.php (Last accessed: March 6, 2013). USDA Forest Service, 2011a. National Report on Sustainable Forests — 2010. Available online at: http://www.fs.fed.us/research/sustain/docs/national-reports/2010/2010sustainability-report.pdf (Last accessed: April 1, 2013). USDA Forest Service, 2011b. Wood supply and demand. Available online at: http://www. fpl.fs.fed.us/documnts/pdf2011/fpl_2011_ince001.pdf (Last accessed: April 1, 2013). Verlegh, P.W.J., Steenkamp, B.E.M., 1999. A review and meta-analysis of country-of-origin research. J. Econ. Psychol. 20 (5), 521–546. Verlegh, P.W.J., Steenkamp, B.E.M., Meulenberg, M., 2005. Country-of-origin effects in consumer processing of advertising claims. Int. J. Res. Mark. 22, 127–139. Vidal, N., Kozak, R., 2009. From forest certification to corporate responsibility: factors influencing adoption and implementation of responsible practices. XIII World Forestry Congress. Bueno Aires, Argentina, 18–23. October, 2009 (Available online at: http:// www.forestry.ubc.ca/sbml/pdfs/SFM,%20CSR%20and%20Forest%20Certification/From_ forest_FD[1].pdf. Last accessed: April 29, 2012). Vriens, M., Wedel, M., Wilms, T., 1996. Metric conjoint segmentation methods: a Monte Carlo comparison. J. Mark. Res. 33 (1), 73–85. Waddock, S.A., Graves, S.B., 1997. The corporate social performance financial performance link. Strateg. Manag. J. 18 (4), 303–319. Wan, Z.F., Geng, Y.D., 2004. The barriers of Chinese state owned forest enterprises reforms. China For. Bus. 64, 13–15 (In Chinese). Wang, L., Juslin, H., 2011. Corporate social responsibility in the Chinese forest industry: understanding multiple stakeholder perceptions. Corp. Soc. Responsib. Environ. Manag. 20 (3), 129–145. Welford, R.J., 2003. Corporate social responsibility in Europe and Asia: critical elements and best practice. J. Corp. Citizenship 13, 31–47. Wilson, N.L.W., 2012. How the cookie crumbles: a case study of gluten-free cookies and random utility. Am. J. Agric. Econ. 94 (2), 576–582. Woodall, C.W., Ince, P.J., Skog, K.E., Aguilar, F.X., Keegan, C.E., Sorenson, C.B., Hodges, D.G., Smith, W.B., 2012. An overview of the forest products sector downturn in the United States. For. Prod. J. 61, 595–603. Xiang, C.M., 2001. Comparison of different survey methods. Shanghai Stat. 12, 19–20 (In Chinese). Zhang, J., Gan, J., 2007. Who will meet China's import demand for forest products? World Dev. 35, 2150–2160.