3.2 Data collection
This study selects Jiangsu Province, China as the research area (see Fig. 1). Jiangsu Province is a major poultry meat production province in China. China's well-known breeding enterprises such as Wen's and Lihua, developing broiler breeding industry in Jiangsu area and leading local farmers to raise chickens and become rich. In addition to these leading enterprises, there are many broiler enterprises with different contract models in Jiangsu. The cooperation models between enterprises and farmers are different, which provides a good research sample for contract difference analysis. The contract arrangements are different, which provides a good research sample for empirical analysis in this paper.
The research team conducted surveys in 11 counties in Jiangsu from March to June 2016, and randomly selected 35 large-scale farmers in each county to conduct a one-to-one survey. The investigator interviewed the head of each sample farmer for 1 hour to obtain the basic characteristics, farming characteristics, production status, contract relationship, and other information about the farmer. Each questionnaire farmer was offered 20 yuan to compensate for loss of work time. In addition, the research team also conducted risk preference and time preference experiments and paid farmers cash in accordance with the results of the experiment. On average, each farmer could receive 50 yuan.[1] A total of 385 valid questionnaires were received, of which 345 were from farmers who farmed broilers with yellow feathers. The specific distribution of sample farmers is as shown in Table 1.
Table 1 Distribution of sample areas
|
Southern Jiangsu
|
Central Jiangsu
|
Northern Jiangsu
|
Total
|
|
Suzhou, Changzhou, Zhenjiang
|
Taizhou, Nantong, Yangzhou
|
Suqian, Xuzhou, Lianyungang, Yancheng, Huaian
|
11
|
|
Taicang, Jintan, Dantu
|
Jiangyan, Haimen, Gaoyou
|
Suyu, Pizhou, Donghai, Dafeng, Xuyi
|
11
|
Observations
|
127
|
64
|
154
|
345
|
3.3 Specification of variables
3.3.1 Explained variable
This paper takes the broiler breeding process as an example. Using whether the farmers have special sewage treatment equipment or devices, waste (feed bags, veterinary drug vaccine packaging), sick chickens, and feces measure the cleaner production behavior of farmers. Through cluster analysis, the waste treatment methods of farmers are divided into cleaner production and non-cleaner production[2].
3.3.2 Key explanatory variable
Contract farming. The contract farmers in this paper refer to the farmers who have signed production contracts with leading enterprises. In this situation, the value of C is 1. The non-contract farmers are the ones who do not participate in the model. So, the value of C is 0. Farmers participating in the production contracts can obtain technical guidance from the enterprise in time. Otherwise, the enterprise will conduct farming norms and requirements for the farmer and supervise the implementation of cleaner production (Acemoglu et al., 2007; Mao et al., 2019). Therefore, farmers who have signed contracts with enterprises are more inclined to cleaner production.
Contract terms. Contract terms mainly refer to the time (unit, month) for signing a single contract between the farmer and the contracting enterprise, which generally includes three months, twelve months, thirty-six months, and one hundred and twenty months. The length of the contract period determines the farmers' behavior expectation, The short-term or uncertain contract term will induce the farmers' short-term behavior in production. The longer the contract period is, the longer the farmers have been engaged in production in the industry. Meanwhile, the higher sunk cost they face, the more incentive they will have to adopt cleaner production technology. In order to save transaction costs, enterprises are more willing to provide more technical training and guidance for long-term contracted farmers to promote cleaner production (Gramzow et al., 2018; Mao et al., 2019).
Social trust. Due to the difference in the objects of interpersonal trust and institutional trust, the mechanism of impact on farmers' cleaner production is different (Braganza et al., 2020). This article studies the impact of interpersonal trust and institutional trust on farmers' cleaner production behavior. Trust in relatives and friends (trust1) represents interpersonal trust. Trust in government (trust2) describes institutional trust. Use the following questions to measure: “Do you think your relatives and friends are trustworthy?”, “Do you think the local government is trustworthy?”. Both questions are ranked in terms of trust intensity (very distrustful, not too trustworthy, generally trustworthy, somewhat trustworthy, very trustworthy).
Table 2 reports the impact of contracts and social trust on farmers' cleaner production behavior. In Table 2, it can be seen that the higher the trust of farmers in relatives and friends and the government, the higher the probability of cleaner production behavior.
Table 2 The impact of contracts and social trust on farmers' cleaner production behaviors
|
|
|
Contract farmer
|
Non-contract farmer
|
|
|
Level of trust
|
Cleaner production(%)
54
|
Non-cleaner production(%)
230
|
Cleaner production(%)
6
|
Non-cleaner production(%)
55
|
Interpersonal trust
|
trust1
|
Very distrustful
|
0
|
0
|
16.67
|
43.64
|
Not too trustworthy
|
11.11
|
55.65
|
33.33
|
10.91
|
General trustworthy
|
18.52
|
11.74
|
16.67
|
36.36
|
Somewhat trustworthy
|
22.22
|
25.65
|
16.67
|
9.09
|
Very trustworthy
|
48.15
|
6.96
|
16.66
|
0
|
Institutional
trust
|
trust2
|
Very distrustful
|
1.85
|
6.96
|
0
|
9.09
|
Not too trustworthy
|
7.41
|
15.65
|
0
|
52.73
|
General trustworthy
|
9.26
|
38.70
|
50.00
|
21.82
|
Somewhat trustworthy
|
35.19
|
29.13
|
33.33
|
16.36
|
Very trustworthy
|
46.30
|
9.57
|
16.67
|
0
|
3.3.3 Control variable
Referring to similar studies (Hoff and Soares-filho, 2021; Mao et al., 2019; Tama et al., 2021; Zhou et al., 2020), this paper controls other variables that may affect the cleaner production behavior of farmers, including the farmers' basic characteristics, farming characteristics and regional characteristics.
Individual characteristic variables: (1) Age. The older the decision-makers are, the more inclined to adopt the previous experience, and the lower the ability to adopt cleaner production technology. Therefore, the older decision-makers are less likely to adopt cleaner production technology (Bellemare, 2012). (2) The educational level of the decision-makers. Owe to advanced thought and broad horizon, the decision-makers with higher educational levels have stronger ability to accept new knowledge and are more inclined to cleaner production (Key, 2003).
Family characteristic variables: (1) The family with more assets is wealthier and more capable of cleaner production (Conley et al., 2010). (2) Regional variables. According to the economic development and climate conditions, the survey area is divided into the northern Jiangsu area which is assigned a value of 1 and the non-northern Jiangsu area which is assigned a value of 0.
Farming characteristics: (1) Farming experience. The farming experience is expressed by the breeding experience of broilers. Farmers with longer breeding experience are more experienced and more likely to be subject to traditional thinking habits. They are used to using the previous breeding mode and less possible to use cleaner production technology (Mao et al., 2019). (2) The scale of broiler farming. For the sake of scale efficiency, the farmer who has larger scale of broiler breeding are more willing to improve the breeding environment through cleaner production (Bojnec and Latruffe, 2013). (3) Breeding technology exchange. Technology exchange can improve farmers' understanding of technology and promote farmers' cleaner production behaviors (Krishnan et al., 2014).
Preference characteristics: (1) Risk preference. Farmers who have a higher degree of risk avoidance are often hesitant to try new things and dare not accept them rashly. In order to avoid the risk of capital investment, they often use previous experience and are less willing to cleaner production. Based on Tanaka et al. (2010) and Liu (2013), the study uses experimental economics methods to measure farmers' risk preferences. Furthermore, by considering expected utility theory and prospect theory, we design experiments to measures farmers' risk aversion. 1-σ is used to describe the curvature of the value function. It can also be used to describe the degree of risk aversion. The larger value of σ indicates that farmers have a higher degree of risk aversion and are less willing to take risks and adopt cleaner production technology. (2) Time preference. Farmers show heterogeneity not only in risk preferences but also in time preferences. Based on the time preference experiments of Tanaka et al. (2010) and Duflo et al. (2011), we calculate the discount rate factor of farmers. The farmers are classified into “short-sighted” farmers and “far-sighted” farmers according to the time preference heterogeneity. Far-sighted farmers have a higher degree of patience, which is more conducive to farmers' cleaner production behavior. The descriptive statistics of variables are shown in Table 3.
Table 3 Variable definitions and descriptive statistics
Variable
|
Definition
|
mean
|
Standard Deviation
|
Explained variable
|
|
|
|
Cleaner production adoption dummy
|
Equal to 1 if the farmer had adopted the cleaner production; otherwise, equal to 0
|
0.174
|
0.380
|
Key explanatory variable
|
|
|
|
Contract adoption dummy
|
Equal to 1 if the farmer had participated in the contract, otherwise, equal to 0.
|
0.823
|
0.382
|
Interpersonal trust
|
Degree of trust in relatives and friends
|
2.991
|
1.165
|
Institutional trust
|
Degree of trust in government
|
3.164
|
1.006
|
Long-term contract adoption dummy
|
Equal to 1 if the farmer had participated in a contract for 1 year or more; otherwise, equal to 0.
|
0.313
|
0.464
|
Control variable
|
|
|
|
Age
|
Actual age of household head in 2015
|
49.342
|
9.578
|
Education
|
Years of formal education of the household head
|
7.238
|
2.797
|
Experience
|
Number of broiler breeding experience (year)
|
6.859
|
4.509
|
Scale
|
Number of broiler breeding scale (ten thousand)
|
4.605
|
4.048
|
Family assets
|
Number of family cars
|
0.414
|
0.609
|
|
Number of air conditioners
|
1.725
|
1.199
|
|
House area(100m2)
|
2.429
|
1.517
|
Technology exchange
|
Equal to 1 if the farmer had no exchange with other farmers;
Equal to 2 if the farmer had very little exchange with other farmers;
Equal to 3 if the farmer had general exchange with other farmers;
Equal to 4 if the farmer had much more exchange with other farmers;
Equal to 5 if the farmer had a lot of exchange with other farmers
|
4.043
|
0.944
|
Region
|
Equal to 1 northern Jiangsu area; equal to 0 non-northern Jiangsu area
|
0.446
|
0.498
|
Instrument variable
|
|
|
|
Distance
|
Distance to the nearest breeding enterprise(km)
|
20.020
|
11.531
|
Table 4 shows the basic situation of the surveyed farmers. The data shows that the majority of the sample farmers are between 41 and 55 years old, which indicates that the age of the decision-maker is not very old. The years of education are mostly 8 years or less, which indicates that the overall education level of the farmers is not high. And most of the farming experience are 8 years or less, which shows that farmers have less experience in breeding.
Table 4 Variable definitions and descriptive statistics
item
|
Proportion (%)
|
item
|
Proportion (%)
|
Age
|
|
Years of formal education
|
|
Under 40
|
18
|
5 years and below
|
27
|
41 to 55 years old
|
56
|
6 to 8 years (including 8 years)
|
54
|
56 to 65 years old
|
22
|
9 to 11 years (including 11 years)
|
18
|
66 years old and above
|
4
|
More than 12 years
|
1
|
Region
|
|
Breeding experience
|
|
South Jiangsu
|
44
|
5 years and below
|
45
|
Central Jiangsu
|
19
|
6 to 8 years (including 8 years)
|
23
|
South Jiangsu
|
37
|
9 to 11 years (including 11 years)
|
18
|
|
|
More than 12 years
|
14
|
Table 5 shows the cleaner production behavior of the surveyed farmers. The data shows that among the 345 broiler farms, there are 21 farmers with special sewage treatment equipment or devices[3], accounting for 6.09%. The waste (veterinary drug vaccine packaging) is selected for sale and reuse can be considered as cleaner production. There are 36 farmers who choose to sell or reuse, accounting for 10.43%; and 309 farmers who choose to discard or incinerate, accounting for 89.57%. Saling and reusing the waste (feed-bag) can be considered as cleaner production. Among them, 250 farmers choose to sell, accounting for 72.46%; 13 farmers choose to discard, accounting for 3.77%; 20 farmers choose to incinerate, accounting for 5.80%; 62 farmers choose other methods such as recycling, accounting for 17.97%. For the treatment of sick and dead chickens, the way to bury deeply can be considered as cleaner production. Among them, 105 farmers choose to sell, accounting for 30.43%; 152 farmers choose to bury deeply, accounting for 44.06%; 3 farmers choose septic tank compost, accounting for 0.87%; 85 farmers choose to incinerate, accounting for 24.64%. The choice of processing organic fertilizer as the method of excrement treatment can be regarded as cleaner production. Among them, there is 1 farmer who chooses to discard, accounting for 0.29%; 14 farmers who choose to return directly to the field, accounting for 4.06%; 5 farmers choose septic tank compost, accounting for 1.45%; 325 farmers choose to process organic fertilizer, accounting for 94.29%.
Table 5 The basic situation of farmers' cleaner production behaviors
Project
|
Proportion (%)
|
Project
|
Proportion(%)
|
Sewage treatment equipment
|
|
Treatment of sick and dead chickens
|
|
Yes
|
6.09
|
Sell
|
30.43
|
No
|
93.91
|
Deeply bury
|
44.06
|
Treatment of veterinary drug vaccine packaging
|
|
Septic tank compost
|
0.87
|
Sell or reuse
|
10.43
|
Incinerate
|
24.64
|
Discard or incinerate
|
89.57
|
|
|
Feed bag treatment
|
|
Feces treatment method
|
|
Sell
|
72.46
|
Unprocessed
|
4.35
|
Discard
|
3.77
|
Septic tank compost
|
1.45
|
Incinerate
|
5.80
|
Process organic fertilizer
|
94.20
|
recycle
|
17.97
|
|
|
[1] The interviews lasted 1 hour per farmer. The total of the farmer’s lost work subsidy and the bonus of the lottery game are equivalent to the local 0.5-day working income. The number of subsidies in this paper is reasonable and has a certain incentive nature.
[2] Cleaner production behavior: The four types of cleaner production activities of farmers are as follows: whether to have special sewage treatment equipment or devices, waste bags (feed, veterinary drugs, vaccines packing materials), dead chicken and waste handling, the four types of waste disposal. By k-means clustering analysis, farmers can be divided into 2 groups, the cleaner production and the non-cleaner production.
[3] There are special sewage treatment equipment or devices can be regarded as cleaner production