Assessments of Landowners’ Willingness to Accept Compensation for Participating in Forest Certiﬁcation in Shandong, China

: Achieving sustainable management of forests in China is becoming increasingly important with more awareness and realization of the importance of forests in environmental protection. Forest certiﬁcation, a market-based instrument to promote sustainable forest management, has been recognized by many countries including China. While landowners’ perception and perspective regarding this voluntary program have been well-documented in the literature, how to motivate and incen-tivize landowners to participate in forest certiﬁcation remained understudied questions. With the face-to-face survey of landowners in Shandong, China, this study analyzed landowners’ willingness to accept compensation for participating in forest certiﬁcation. Results indicated that the average accepted compensation in terms of increased timber price was about RMB 120 (~USD 17.6)/m 3 to have their forests enrolled in certiﬁcation programs. Results from multiple regression showed that the level of such compensation required for participation in certiﬁcation was inﬂuenced by forestland size, owner demographics, certiﬁcation-program requirements, and importance placed on timber production. These ﬁndings would inform policy-makers in designing compensation-related policies and establish incentive-based mechanisms to motivate forest landowners to participate in forest certiﬁcation programs in China.


Introduction
Achievement of sustainable forest management for environmental, social and economic benefits is the aim of China's modern (after 1990 s) forestry development as natural forests are being depleted because of overharvesting of timber over its history [1][2][3]. In China, example practices of sustainable forest management include restoration of degraded landscape and improvement of forest ecology and biodiversity through government initiatives and policies [3]. Forest certification as a market-based mechanism has been recognized in China in the 1990 s and the function in promoting forest management level has also been affirmed. Therefore, Chinese officials began to explore and develop forest certification systems from 2001 based on China's forestry conditions and thus far, China has three major certification programs including the Forest Stewardship Council (FSC), Programme for the Endorsement of Forest Certification schemes (PEFC), and China Forest Certification Council (CFCC). Both the FSC and PEFC are internationally recognized certification schemes and the difference is that wood processors (in addition to other entities) are more involved in the origin of the PEFC. The CFCC system is a program developed by China itself and is endorsed by the PEFC. Although there are differences in the certification schemes' stan-Sustainability 2021, 13, 903 3 of 15 size of forestland and those who intend to harvest timber shortly were more likely to adopt certification than otherwise [14,17,18]. Furthermore, the research of Bell et al. [23] and Tian et al. [18] reported that landowners who have a management plan for their forestland were more likely to adopt certification than their counterparts, which was ascribed to landowners' previous experience with forest management positively associated with their willingness to participate in forest stewardship programs.
The literature thus far provides a good perspective on landowners' awareness and perceptions of forest certification and their potential concerns/challenges. What is scarce in certification literature is an estimation of the compensation necessary for landowners to certainly participate in a certification program. Since landowners who are considering enrolling in certification may face a small cost of participation (e.g., third-party verification fee, potential harvest reduction, added cost in implementation of best management practices), it may not be financially worthwhile for many without some compensation or increase in timber price. In addition, understanding what specific attributes of forest ownership, forest households' demographics, and requirements of certification program schemes impact the necessary compensation is important in designing incentive-based programs. Therefore, the objectives of this study were to (1) investigate landowners' familiarity and attitudes towards forest certification, (2) examine differences in ownership and forestland characteristics among subgroup landowners, and (3) estimate landowners' average willingness to accept (WTA) compensation for participating in forest certification and identify factors influencing it. Through identifying the characteristics of landowners associated with an accepted WTA compensation for participating in forest certification programs, this study could provide useful information for policy-makers in designing compensation-related policies and establish incentive-based mechanisms to motivate landowners to enroll in certification programs in Shandong and elsewhere.

Data Collection
Data for this study were collected in 2016 from the Shandong province of China using a face-to-face survey (Supplementary Materials). The survey was reviewed and approved on 20 June 2016 by the University of Tennessee's Institutional Review Board (IRB 16-02999-XP) before implementing. Shandong Province is a coastal province located in the East China region. There is a diverse geomorphologic pattern in Shandong, which has mountainous and hilly areas in both the east and west as well as the center; in addition, there are great plain areas in the north west of Shandong which were formed by the alluviation of the Yellow River. Besides, there are also coastal areas which mainly located in the northeast and southeast regions. The diverse and unique geographical location brings about rich plant species and, regarding forest vegetation, has 615 species and is dominated by broad-leaf deciduous forest. In terms of climate, this province is located in a monsoon climate region with a moderate warm temperate monsoon climate. The annual average temperature and precipitation is around 11-14 • C and 550-950 mm, where most of (60-70%) the rainfall precipitation happens in summertime between June and August. To increase timber production and forest cover, the forest sector in this province focuses on expanding both industrial plantations and ecological afforestation since the 1990s. A famous project in this province is the Shandong Ecological Afforestation Project which was implemented during the period of 2010~2015. The industrial timber production forest in Shandong mainly refers to the fast-growing and high-yield poplar forests [17,25]. The forest area in this province is, in total, around 2.55 million hectares which is mainly distributed on the mountainous and hilly areas as well as the coastal areas. The survey conducted in this study were completely in six heavily forested cities including Taian, Jinan, Linyi, Liaocheng, Jining, and Weifang, which are located in the mountainous and coastal areas.
The target population for this survey was forestland households in Shandong Province. The survey was administered under the help of local forest officials, who provided in-formation about who owns forests locally and a random sample was then selected to approach either at their residences or at their forestlands. The dominated tree species among the interviewed landowners is poplar forests. Taking Linyin as an example, the total poplar plantation area is about 200 thousand hectares [6]. The questionnaire was filled by totally 557 landowners (households in China have time-limited (generally between 30 and 70 years) forest use and management rights and this kind of "ownership" is ascribed to China's forest land tenure. To make this term understandable, we use "landowners" in this study for those households who currently have land use and management rights.) and 50 of them did not finish the questions in the survey, so we had 507 usable completed surveys. Survey questions primarily included Likert scale items of various scales (e.g., 1 = not familiar at all, 5 = extremely familiar; 1 = not interested at all, 4 = very interested). We also included an open-ended WTA question asking about landowners' minimal acceptable timber price (RMB/m 3 ) increase (percentage increase from the current price) to definitely participate in a forest-management certification program. Other survey questions included questions regarding ownership and forestland characteristics (e.g., forestland size, tenure, intent to harvest timber, acquisition mode of forestland by landowners), and the likelihood of participating in certification if the program had a given condition of participation and sociodemographic information (e.g., age, gender, income, education, occupation, and living area) about landowners [17,18]. In addition to the survey questions, the secondary data of the poplar timber price were collected.

Econometric Model
We hypothesis that landowners' awareness and knowledge of forest certification might positively correlate with their interest level in adopting a forest certification program; therefore, to test the correlation between landowners' familiarity with forest certification and their interest level in participating in a certification program, Kendall's rank correlation tau was used. Kendall's rank correlation is a traditional correlation method for two ordinal variables [26]. Meanwhile, analysis of variance (ANOVA) was employed to test for differences in subsequent groups of not familiar and familiar groups at a 0.05 significance level. The associated information of forestland and ownership characteristics, as well as landowners' demographics, were tested to further describe the subgroup households.
Following Lindhjem and Mitani [27] and Beach et al. [28], landowners' WTA compensation can conceptually be specified as a function of resource conditions (e.g., forestland condition), ownership characteristics, conservative program (e.g., forest certification) requirements, and market drivers (e.g., interest rate, timber price, etc.). specifically, the function is expressed as (Equation (1)): where: FR is a vector of forestland characteristics such as land size, harvesting plan, etc., OC includes owners' sociodemographic variables such as age, gender, income and education, etc. PR denotes a vector of variables associated with certification program design characteristics such as whether a management plan was required for participating in a forest certification scheme, etc. OT mainly represents other variables such as timber price and interest rate, etc.
Regarding the willingness to pay (WTP)/WTA data collected with an open-ended question, a very traditional analysis method is the ordinary least squares method (OLSs) [27], but this statistical method has a requirement that the data distribution is a peculiar distribution. However, the WTA data gathered in this study are a continuous positive variable indicating landowners' acceptable compensation amount. As a result, the econometric estimation approach employed was the Tobit model [29][30][31][32]. The dependent variable was landowners' acceptable timber price increase (RMB/m 3 ) for participating in a forest management certification program. This regression model was undertaken in the software of Statistical Analysis System (SAS). Landowners' WTA compensation was measured in terms of percentage increase in timber price from the average price of past fifteen years. To be specific, respondents' indicated that the percentage increase in timber price was multiplied by the average timber price at the time of the survey (RMB660 (~USD 96.5)/m 3 ) to quantify the monetary amount of compensation. Independent variables consisted of variables related to ownership and forestland characteristics (e.g., tenure and land size), owner sociodemographic, as well as program design characteristics. Table 1 presents the description, mean and standard deviation (S.D.) for both dependent and potential influential explanatory variables that were used for empirical analysis. The Tobit model is presented as Equation (2): where: x is the matrix of all possible independent variables included in this study, and β represents the vector of coefficients, and ε i is the error term which follows a normal distribution. The group of forest resource characteristics (FR) variables included forestland size, management objective of forestland, acquisition mode of forestland, tenure, timber harvest history, and whether it had a management plan. The management objective related variables primarily concerning the importance of forestland for recreation (e/g., family gathering, scenic view, etc.) and timber production, as well as in protecting nature and biodiversity. The acquisition mode of forestland mainly included the purchase, inherit, and trade/rent options and for each option, we coded it as a dummy variable (Table 1). Timber harvest history consisted of variables of whether landowners harvested timber in the previous five years and whether they intend/plan to harvest timber in the coming five years. The sociodemographic variables (OC) collected in this study primarily consisted of age, gender, education, income, and occupation. Besides, we also considered the effect of landowners' living area which consisted of metropolitan area/suburb area, town/county, and rural/village area, as well as their return from forestland which was measured as the percentage of annual household income from forestland. The specific description for these variables, including definition (i.e., type of variables), is included in Table 1. The third category of CP included different variables in terms of required conditions of certification program. Specifically, respondents were asked to indicate their likelihood of participating in certification program if it had a given condition to participate such as RMNP, if the program requires landowner to have a management plan; ROPI, if the landowner agrees that allowing periodic on-site inspection of forestry practices is a drawback of participation; RPFMH, if the landowner is required to use a professional forester to manage forestland; RICP, whether landowners themselves are required to be involved in some part of the certification progress; and RIRP, if the landowner is required to make the full inspection summary results available to the public. These were measured in a 5-point Likert scale (1, definitely unlikely; 5, definitely likely), but were later converted to a binary dummy variable. This was performed by converting the values of 4 (somewhat likely) and 5 (very likely) into 1, and the remaining values of 1-3 (that represented unsure or unlikely) into 0. In addition, we also considered variables related to landowners' familiarity and interest degree for forest certification (Table 1).

Landowners' Knowledge and Attitudes Towards Forest Certification
Familiarity with forest certification among landowners was very low (Figure 1). Almost 38% of the respondents reported that they had never heard of the concept (not familiar at all) before receiving the survey; about 20% of the respondents indicated that they had heard forest management certification before (slightly familiar) but did not understand its meaning. About 22% of the respondents indicated that they had at least a minimal understanding of forest certification (very familiar), and only 2% of the respondents said that they had an extensive understanding of forest certification (extremely familiar).

Landowners' Knowledge and Attitudes Towards Forest Certification
Familiarity with forest certification among landowners was very low (Figure 1). Almost 38% of the respondents reported that they had never heard of the concept (not familiar at all) before receiving the survey; about 20% of the respondents indicated that they had heard forest management certification before (slightly familiar) but did not understand its meaning. About 22% of the respondents indicated that they had at least a minimal understanding of forest certification (very familiar), and only 2% of the respondents said that they had an extensive understanding of forest certification (extremely familiar).
Even landowners had a low familiarity with forest certification, after giving them a brief description and introduction for forest certification such as the associated benefits for adopting forest certification and the purpose of certification, etc. About 63% of the respondents expressed some interest in participating in forest certification. To test the hypothesis that landowners might be more interested in adopting forest certification if they were more aware of this concept [10], we used Kendall's rank correlation tau to examine their correlation. The correlation result (p = 0.22) suggested that there was no significant correlation between them, so the premise was incorrect.

Characteristics of Subgroup Landowners
We classified the respondents into different subgroups according to their familiarity with the forest certification program ( Table 2). Respondents in the familiar group (n = 117) were those who reported that they understood the basic concepts or general certification process or have been involved in the certification process before. By contrast, the respond-   Even landowners had a low familiarity with forest certification, after giving them a brief description and introduction for forest certification such as the associated benefits for adopting forest certification and the purpose of certification, etc. About 63% of the respondents expressed some interest in participating in forest certification. To test the hypothesis that landowners might be more interested in adopting forest certification if they were more aware of this concept [10], we used Kendall's rank correlation tau to examine their correlation. The correlation result (p = 0.22) suggested that there was no significant correlation between them, so the premise was incorrect.

Characteristics of Subgroup Landowners
We classified the respondents into different subgroups according to their familiarity with the forest certification program ( Table 2). Respondents in the familiar group (n = 117) were those who reported that they understood the basic concepts or general certification process or have been involved in the certification process before. By contrast, the respondents in the not familiar group (n = 388) were those who never heard the concept before or have heard it but did not know/understand what it means before receiving our survey. Results of ANOVA tests indicated that there were typological differences in terms of respondents' socio-demographics and their forestland characteristics between those two subgroups. No significant difference was found in the average age between these two subgroups, but a marked distinction was found in their education level, income level, living area, as well as occupation. Specifically, the percentage of respondents in the familiar group who gained a bachelor/higher degree (76.9%) was significantly higher than the not familiar group (56.2%), while the percentage of who did not complete high school in this group (1.7%) was much lower than the not familiar group (14.9%). In terms of income level, 17% of respondents in the familiar group had an annual income level greater than RMB 75,000, whereas this number was only 6% in the not familiar group. Almost half of (51.9%) the respondents in the not familiar group lives at the rural/village area which was marked higher than the percentage (28.5%) in the familiar group. By examining the respondents' occupation, results suggested that over half of the respondents (57.3%) in the familiar group were professional managers while almost half of the respondents in the not familiar group were farmers (49.7%). In addition, the forestland characteristics were also examined for these two subgroups and no substantial difference was found for tenure, importance of timber production, and whether to harvest timber in the near future (Table 2). Respondents in both groups saw a high importance in timber production and they both intended to harvest timber in the near future. The average forestland size was about 125 hectares in the not familiar group and it was around 192 hectares in the familiar group. In terms of having a management plan for their forestland, a relatively higher percent of respondents in the familiar group (62%) had a management plan than those in the not familiar group (49%). Regarding respondents' future plan of forestland, over two-thirds of respondents in both groups indicated to self-manage it, whereas 16% of respondents in the not familiar group planned to sell/rent out, but only 6% in the familiar group.

Landowners' Accepted WTA and Affecting Factors
Results of the survey data revealed that respondents' accepted minimum timber price increase averaged 18% from the current timber price for them to certainly have their forestland certified. Additionally, the calculated expected/average poplar timber price was around RMB660 (~USD 96.5)/m 3 based on the past 15 years' prices (2005-2020) ( Figure 2); as a result, the accepted timber price increase for respondents to participate in forest certification was around RMB120 (~USD 7.6)/m 3 .
Results of the survey data revealed that respondents' accepted minimum timber price increase averaged 18% from the current timber price for them to certainly have their forestland certified. Additionally, the calculated expected/average poplar timber price was around RMB660 (~USD 96.5)/m 3 based on the past 15 years' prices (2005-2020) ( Figure  2); as a result, the accepted timber price increase for respondents to participate in forest certification was around RMB120 (~USD 7.6)/m 3 .  (Table 3) and results showed that they were less than the critical threshold of 10 [33], indicating that multicollinearity was not a concern in this model. Initially considered values but exhibited VIF value higher than 10 (i.e., variables under conditions of participation: whether participation would require thorough involvement in the certification process or making part of the summary inspection results available to the public) were removed from the final regression model. Tobit regression results were summarized in Table 3. The results revealed variables in the sociodemographic group-age was not significantly related to respondents' WTA compensation, but gender (p < 0.10) had a positive and significant association with their WTA compensation for participating in a forest certification program. This result implies that male respondents requested a higher WTA compensation than their female counterparts. A similar result was found for the education variable of did not complete high school (p < 0.01). This positive result indicates that respondents who did not complete high school ask for a higher WTA compensation than those who at least have a high school education in terms of adopting forest certification. Variables of income and the percentage of household income comes from forestland were not significantly associated with respondents' WTA compensation. For variables related to occupation, a positive and significant relationship was found for the occupation variable of government employee (p < 0.10), which implies that WTA goes up if a respondent is a government employee. As expected, the variable of rural/village in terms of living area (p < 0.001) had a significant and positive association with respondents' acceptable compensation. This positive coefficient indicates that respondents who are living in a rural/village area demand more WTA compensation to have their forestland certified.

Multicollinearity was examined by calculating the index of variance inflation factors (VIFs) among all independent variables
The results of the forestland and ownership characteristics group indicated that forestland size was significantly (p < 0.001) and negatively related to a respondent's WTA  (Table 3) and results showed that they were less than the critical threshold of 10 [33], indicating that multicollinearity was not a concern in this model. Initially considered values but exhibited VIF value higher than 10 (i.e., variables under conditions of participation: whether participation would require thorough involvement in the certification process or making part of the summary inspection results available to the public) were removed from the final regression model. Tobit regression results were summarized in Table 3. The results revealed variables in the sociodemographic group-age was not significantly related to respondents' WTA compensation, but gender (p < 0.10) had a positive and significant association with their WTA compensation for participating in a forest certification program. This result implies that male respondents requested a higher WTA compensation than their female counterparts. A similar result was found for the education variable of did not complete high school (p < 0.01). This positive result indicates that respondents who did not complete high school ask for a higher WTA compensation than those who at least have a high school education in terms of adopting forest certification. Variables of income and the percentage of household income comes from forestland were not significantly associated with respondents' WTA compensation. For variables related to occupation, a positive and significant relationship was found for the occupation variable of government employee (p < 0.10), which implies that WTA goes up if a respondent is a government employee. As expected, the variable of rural/village in terms of living area (p < 0.001) had a significant and positive association with respondents' acceptable compensation. This positive coefficient indicates that respondents who are living in a rural/village area demand more WTA compensation to have their forestland certified.

Multicollinearity was examined by calculating the index of variance inflation factors (VIFs) among all independent variables
The results of the forestland and ownership characteristics group indicated that forestland size was significantly (p < 0.001) and negatively related to a respondent's WTA compensation for certifying their forestland. This result suggests that the WTA decreases with respondents' land size. Regarding the forestland management objective variable of the importance of managing forestland for protecting biodiversity, there was a significant and positive (p < 0.05) correlation with respondents' WTA. A similar positive and significant result was observed for the dummy variable representing whether respondents have a management plan for their forestland (p < 0.01). This observation implies that respondents' WTA is higher for those who have a management plan than those who do not have one. No significant association was found for other variables in this category including the acquisition mode of forestland, tenure, as well as future harvesting plan. The certification program design characteristics category included different requirements of the certification program after participating, specifically, the requirements of having a management plan, on-site inspection, having a professional forester to manage and harvest forestland, as well as involvement of the certification process and publishing summary inspection results to the public. Among these variables, results revealed that there was a significantly positive (p < 0.10) association between the requirement of having a management plan with respondents accepted compensation for adopting forest certification. A similar result was found for the requirement of using professional forester/loggers (p < 0.01) to harvest forestland. Both results suggest that respondents' WTA increases with the requirements of having a management plan and having trained foresters/loggers to harvest timber for participating in a forest certification program.
Regarding the variable in the group of market drivers, results suggested that the economic importance of timber sales was not significantly and positively associated with respondents' WTA. Likewise, with regard to the variables describing respondents' familiarity and interests in forest certification, results in this research revealed that there was no significant association between them with respondents' WTA.

Discussion
The survey results of this study indicated that landowners in Shandong Province have a generally low familiarity with forest certification, which was consistent with the previous associated studies [15,17,18,34]. The study of Chen et al. [34] found that, in China, wood product manufacturing was not familiar with forest certification either. The correlation results suggested that there was no significant association between landowners' familiarity with forest certification and their willingness to participate in this program. This was not in line with the studies of Mercker and Hodges [35] and Sun et al. [36], who reported that there was a positive association between landowners' knowledge of forest certification and their likelihood of participating.
Significant differences were found in landowners' demographics between the familiar and not familiar group. In terms of education and household income, the familiar group landowners were more educated and had a higher income level than those in the not familiar group, which was consistent with He et al. [16], who stated that education was positively associated with landowners' familiarity with forest conservation programs. In addition, results indicated that occupation was associated with landowners' familiarity with forest certification since almost two-thirds of landowners in the familiar group worked as foresters or professional managers. Regarding ownership variables, we found a significant difference in ownership size between these two subgroups, suggesting a positive association between ownership size with a landowner's familiarity with forest certification. Between the two subgroups, a significant difference was found in the availability of a written forest management plan, which implied a positive correlation between having a management plan and landowners' familiarly with forest certification.
The ballpark estimation for landowners' accepted WTA to participate in a forest management certification program was around RMB120 (USD 17.6)/m 3 in Shandong Province. This estimated amount was very close to a local city of Linyi's news report [6], which showed that the price for certified poplar forest products is averagely 20% higher than those not certified. This could bring about extra income of RMB100 (~USD 14.6)/m 3 for landowners [6]. The RMB20 (~USD 2.9)/m 3 difference resulted from the average price difference; specifically, the RMB120 (~USD 17.6)/m 3 was estimated based on the average timber price of the past 15 years while in the news it only used the average timber price of 2013.
Regression results of this study displayed that a landowner's minimum accepted compensation for participating in a forest certification program was influenced by a series of ownership and forestland characteristics. Among the sociodemographic variables, a positive and significant coefficient was found for gender, indicating that male landowners require a higher WTA compensation than their female counterparts for adopting a forest certification program. Research [20][21][22] found that women are more concerned with environmental issues than men, which possibly relates to the higher WTA for male landowners. A positive and significant coefficient was found for the variable of education, which implies that if a forest household's highest education level is not finishing high school, he/she would demand an RMB28 (~USD 4.1) higher compensation for participating in forest certification than those who at least finish high school. Likewise, a positive and significant effect was found for the variable of government employee, which suggests that if a landowner is/was a government employee, he/she asks a higher WTA compensation for having his/her forestland certified than others. Similarly, those who lived in a rural/village area requested more compensation for them to have their forestland certified, which is possibly because landowners in this area are more financially reasoned to their own forestland. In addition, the variable of income was not significantly associated with landowners requested compensation to seek forest certification. This finding was in line with Kline et al. [20], who stated that willingness to accept compensation for landowners to participate in forest conservation program was quite consistent among different income categories.
A positive coefficient before the variable of size indicated that a landowner who owns a large acreage size of forestland requires less compensation to definitely have their forestland certified. This observation is not particularly surprising considering that the average per unit certification cost would be lower for a large forestland size than a small size of forestland, and this result was consistent with studies by Kilgore et al. [19] and Bateman et al. [37]. As expected, forest owners who indicated a high importance of their forestland in protecting nature and biodiversity demand a higher compensation to enroll in forest certification. Landowners would expect a higher compensation amount to pay off their benefits reduction from timber activities for providing the public benefits such as protecting biodiversity. Regarding whether landowners have a management plan for their forestland, a positive and significant coefficient was observed, which implies that landowners who have a management plan would request a higher WTA compensation than those who do not have one. This observation might due to the fact that having a management plan could increase their management costs indirectly.
A positive and significant association was found between different requirements of the forest certification program and landowners' WTA compensation for adoption. Specially, if the certification program request landowners to use a professional forester to manage and harvest their forestland after certification, landowners would demand a higher WTA compensation than otherwise. Moreover, landowners indicated a higher WTA compensation if the certification program requires them to have a written management plan after participation. A possible reason behind those two observations is that hiring a professional forester or having a written management plan might indirectly increase their management costs.

Conclusions
Overall, this study sheds some light on the analysis of landowners' preference and willingness to accepted compensation for adopting forest certification in Shandong Province, China. Compensation needed for landowners, in terms of increased timber revenue, to consider participating in the certification program was estimated at RMB120 (~USD 17.6)/m 3 to have their forestland certified. In other words, it may be the case that landowners might not find it financially worthwhile to participate unless the timber price increases by this amount. This finding may be helpful to government agencies and non-government institutions that are currently considering designing incentive mechanisms to encourage landowners' participation. Moreover, this compensation amount was influenced by a group of ownership and forestland characteristics. For example, to encourage/motivate landowners participating in forest certification, a greater/higher compensation might be required for those who are living in a rural/village area and those who have a low education level such as people who did not finish high school. On the contrary, a lower compensation might be enough for landowners with larger acreage size. In addition, landowners whose forests provide biodiversity protection or recreational uses expected a higher compensation to get certified. There are existing mechanisms that can be implemented by landowners to get financial compensation for such biodiversity conservation and recreational services. One typical example is the implementation of the Ecosystem Services Procedure (ESP) from the FSC scheme. This ESP program could be implemented together with forest certification, which could be helpful in motivating landowners to participate in a certification program. Moreover, another important incentive-based program might be the creation of a carbon credits program which might also bring financial benefits for landowners who value a high importance of biodiversity protection.
In addition, a few caveats about the data analysis and modeling in this study should be noted. First, landowners' willingness to accept compensation was computed by averaging 15 years' poplar timber price which is changeable and uncertain regarding the market condition. The other options could have been direct subsidy or payment from government or other entities to participating landowners. However, in the absence of such programs, increase in timber revenue through higher timber price was used instead to elicit willingness to accept value. Admittedly, this measure was not the same as value elicitation question typically using in soliciting WTA but is a reasonable alternative and perhaps free from hypothetical bias associated with accepting or rejecting a hypothetical bid. Second, some of the variables originally measured on a 5-point Likert scale were converted into binary dummy variables because of the better fit in the regression model. While it is common to adopt such a convention in empirical analysis, its impact on the results is unknown and cannot be completely ruled out.
Future studies could focus on identifying landowners' desirability to learn about forest certification and their preferred and accepted learning methods (e.g., workshop, webinar, or video conference, etc.) would be needed to improve landowners' awareness and familiarity with forest certification in China.