Adoption of voluntary water-pollution reduction technologies and water quality perception among Danish farmers

Agricultural Water Management 158 (2015) 235–244

Contents lists available at ScienceDirect

Agricultural Water Management journal homepage: www.elsevier.com/locate/agwat

Adoption of voluntary water-pollution reduction technologies and water quality perception among Danish farmers Florence Gathoni Gachango ∗ , Laura Mørch Andersen, Søren Marcus Pedersen Department of Food and Resource Economics, University of Copenhagen, Denmark

a r t i c l e

i n f o

Article history: Received 30 April 2014 Accepted 23 April 2015 Available online 26 May 2015 Keywords: Agricultural water pollution Water quality Agri-environmental measures Recipient water bodies Ordered probit

a b s t r a c t The adoption of voluntary nutrient reduction technologies among Danish farmers is relatively low despite the introduction of a number of incentives to do so. With data from 267 farmers, this study analyzes the level of adoption of these technologies and the farmers’ perception of water quality, existing regulatory measures and their implementation strategies. In general, farmers perceive the water quality to be above average and indicate a strong opposition to penalties for non-compliance. Results of two ordered probit models on adoption and perception show a significant importance of farm and soil types, farm size and slopes and information availability. These findings point to the need for increased information dissemination on water quality requirements both at national and regional levels and technical and institutional support for the existing and future incentives. © 2015 Elsevier B.V. All rights reserved.

1. Introduction 1.1. Nutrient reduction plans For over 25 years, Denmark has implemented different approaches to the reduction of nitrogen and phosphorous discharges from agricultural farms. These initiatives, which range from the initial Action Plan for the Aquatic Environment (APAE) in 1987 to the broader Green Growth Agreement in 2009, have mainly been implemented using a national-wide approach as opposed to the designation of vulnerable zones adopted in other EU countries (Smith et al., 2007a). The first APAE, effective in 1987 was followed shortly by APAE II which was implemented in 1998 with the aim of reducing Nitrogen (N) and Phosphorus (P) losses to the aquatic environment by 50% and 80% respectively. The third APAE became effective in 2004 and further targeted N and P reduction by 13% and 50%, respectively, by 2015. An evaluation of the plan in 2008 indicated that only an insignificant decrease in nitrate leaching had been achieved between 2003 and 2007, thus leading to the launch of the Green Growth Agreement (GGA, 2009–2015). The aim of GGA was to integrate activities aimed at implementing and achieving the requirements of the Water Framework Directive (WFD), deal with the problems encountered

∗ Corresponding author at: Department of Food and Resource Economics, University of Copenhagen, 25 Rølighedsvej 1958, Frederiksberg, Copenhagen, Denmark. Tel.: +45 35336831; fax: +45 35336801. E-mail address: [email protected] (F.G. Gachango). http://dx.doi.org/10.1016/j.agwat.2015.04.014 0378-3774/© 2015 Elsevier B.V. All rights reserved.

in APAE III as well as ensure a balance between nature, environment and agricultural development. The targets of the GGA are to reduce nitrogen and phosphorous leaching to coastal waters by 19,000 tonnes and 210 tonnes respectively (Danish Ministry of Environment, 2009). The GGA, aims to ensure environmentally friendly agricultural production by e.g., ensuring sustainable use of resources, stimulating green energy production and promoting market based organic production. The implementation of these initiatives is largely through various government-funded interventions such as the promotion of biogas production, perennial crop production, organic production and the establishment of wetlands. In addition, the compulsory measures developed in the previous action plans are still effective. There are, however, a few exceptions if the producers adopt some of the measures proposed under the GGA although there are no clear guidelines on the exemptions. Table 1 summarizes the nutrient reduction measures in the Danish context (Balticdeal, 2011). Despite the incentives given under the GGA, the adoption and implementation of the proposed measures has been relatively low. This aspect is clearly evident with over 90% of the Danish farms still practicing conventional farming (Statistics Denmark, 2014). Additionally fewer than 100 applications for the construction of biogas plants have been submitted for funding (Jacobsen et al., 2013), while the production of perennial energy crops currently stands at 4000 ha (Ministry of Food, Agriculture and Fisheries, 2008). The construction of wetlands is also very low with less than 20 wetlands covering approximately 130 ha being fully established (supremtech.dk). The area of wetlands is quite small compared to

236

F.G. Gachango et al. / Agricultural Water Management 158 (2015) 235–244

Table 1 Summary of nutrient reduction measures in Denmark. General Measure

Description of specific measures

Non-season plant cover

- Planting of cover crops in winter - Growing of Catch crop - Setting aside of buffer zones - Cultivation of permanent grass

Obligatory by law √ √ √

Tillage

- No tillage in Autumn before spring sown crops

Fertilization

- Compulsory fertilizer plans and accounts - Maximum Nitrogen quotas at farm level (calculation of quotas based on crops, soil types, etc., and minimum use of Nitrogen in animal manure)

Manure/slurry management

- Restrictions on application times and techniques for solid and liquid manure - Separation of liquid–solid manure/slurry - Acidification of manure/slurry - Establishment of manure/slurry storage facilities - Biogas production

Run-off water treatment

Production system

Subsidized through agri-environmental scheme

Future possibility √

√ √ √

√ (On bare soils and grasslands)

√ √

√ √ √

- Wetlands and sedimentation ponds

- Maximum livestock units (Lu) per ha (1.7 Lu/ha for dairy cows and 1.4 Lu/haa for other livestock) - Maximum Phosphorous amounts in livestock feeds - Extensive farming in sensitive areas - Organic production

(in some regions) √ √ √ √

Adapted from http://www.balticdeal.eu/news/new-measures-in-denmark/ (accessed 30.03.14 a 1 cow = 1.33 Lu, 36 pigs @ 32–107 kg = 1 Lu, 200 piglets @ 7.2–32 kg = 1 Lu, 4.3 sows = 1 Lu and 2900 chicken @ 40 days = 1 Lu.

the stated potential of 15,000 ha and the increasing demand for more targeted nitrogen and phosphorous reduction measures in Denmark (Balticdeal, 2011; Hoffmann et al., 2011). Although the low pace of adoption of all these technologies (specifically the constructed wetlands) may be attributed to the fact that their actual effectiveness and/or profitability has not yet been fully established, the phenomenon is not unique to Denmark. Similar low adoption rates of best management practices (BMPs) have been observed in other countries, despite farmers having full information regarding the performance and profitability of these environmental measures (Smith et al., 2007a). This phenomenon therefore calls for a critical assessment of factors influencing the adoption of voluntary pollution mitigating technologies, farmers’ attitudes and perceptions of the quality of surface water, the various nutrient reduction measures and their impact on water quality and the relationship between technology adoption and water quality perception. It is expected that this knowledge and understanding will help policy makers formulate strategies for implementing environmental measures that reduce pollution from agricultural activities. Currently in Denmark, the construction of wetlands is still at a preliminary stage in the form of pilot projects to test the technique and its effects (Ministry of Food, Agriculture and Fisheries, 2012). The initiative is supported through the Danish rural development fund for non-productive investment in agriculture and grants for investment in “new green technologies” under the GGA (Minivådområder, 2012). Constructed wetlands (CWs) are however seen as a more targeted and cost effective option in the reduction of N and P pollution from agricultural fields (Kjærgaard et al., 2012). Consequently, given the expected potential of the technology and the projected capacity of one hectare of wetland to remove 480–1380 kg N per year (Kjærgaard and Hoffmann, 2013), this measure, if properly implemented could effectively replace some of the existing mandatory measures that have direct negative effects on the productivity and profitability of farms.

In order to identify the best strategy for policy makers to incorporate CWs as a nutrient reduction measure and encourage farmers to adopt them, it is paramount to first establish the farmers’ level of adoption of voluntary technologies, their attitudes and perception regarding the current surface water quality, their perceived effectiveness of the existing regulatory nutrients mitigation measures and their preference for various government strategies for implementing the pollution reduction measures. These four aspects and their interrelations are analyzed in this article. 1.2. Previous studies and conceptual model Studies on farmers’ perception of water quality and pollution reduction measures and adoption of associated BMPs have been conducted over the last few decades with most of them being conducted in the US (Lichtenberg and Lessley, 1992; Ryan et al., 2003; Morton, 2007; Popp et al., 2007; Kaplowitz and Witter, 2008; Hu and Morton, 2011; Savage and Ribaudo, 2013). Bratt (2002), analyzes Swedish farmers’ choices for management practices aimed at reducing nutrient pollution at the catchment level, while Sang (2008) studies farmers’ preference for catchment management practices in Scotland. Other studies have been carried out in developing countries (Nguyen et al., 2006; Mojo et al., 2010; Perez-Espejo et al., 2011). Some of these studies have focused on catchment level analysis while others have primarily targeted areas designated as Nitrate Vulnerable Zones (NVZs). Macgregor and Warren (2006) assess arable farmers’ perspectives about the causes of water pollution in two NVZs in Scotland, while Barnes et al. (2009) analyze perceptions about NVZs regulations among farmers with different farm typologies in all the four Scottish NVZs. In general, most of these studies find that farmers acknowledge the existence of water quality problems with agricultural production being a possible source. However, they seem to point out that most farmers do not agree that water pollution results from their own farms. The studies also reveal that farmers are generally opposed to the

F.G. Gachango et al. / Agricultural Water Management 158 (2015) 235–244

237

Fig. 1. Conceptual framework. Adapted from Ervin and Ervin (1982).

adoption of measures and practices that interfere with their production practices. In Denmark, only a few studies on perceptions of water quality and pollution reduction measures have been conducted. A study by Smith et al. (2007b) compares the perspectives of respondents from significant actor groups involved in the process of implementing the WFD in Denmark and England, while Atkins and Burdon (2006) analyze individuals’ preferences for water quality improvements in Randers Fjord located in the Central region of Denmark. Additionally, Christensen et al. (2011) analyze the determinants of farmers’ willingness to participate in subsidy schemes in the implementation of pesticide free buffer zones. Although the study by Smith et al. (2007b) deals with a broader issue, producers’ perspectives are not clearly captured since farmers do not form part of the respondents. Likewise, the study by Atkins and Burdon (2006) does not focus on farmers specifically, but rather targets the general public within the Randers Fjord. Even though the Christensen et al. (2011) study focuses on the farmers’ perspectives, its main emphasis is on their preference for a subsidy scheme in the reduction of pesticide-based pollution. The current study however focuses on nutrients pollution and targets farmers within different catchment areas. Different approaches have been used in perception and adoption studies. The adoption of conservation technologies has mainly been based on the initial work of Ervin and Ervin (1982), and Just and Zilberman (1983), who base adoption on the expected utility theory. More studies have followed over the years and summaries of factors influencing adoption have been reported by Kabii and Horwitz (2006), and Knowler and Bradshaw (2007). Behavioral approaches based on the theory of reasoned action (Ajzen and Fishbein, 1975) and the theory of planned behavior (Ajzen, 1985) have also been employed in many studies (Beedell and Rehman, 2000; Ahnström et al., 2009; Wauters et al., 2010). Strict application of either economic or sociological approaches excludes important factors and therefore a framework that accommodates the two approaches should be used (Mbaga-Semgalawe and Folmer, 2000). In this study, we follow the conceptual framework of the decision making process in the adoption of agricultural technologies proposed by Ervin and Ervin (1982). Although the literature on the adoption of agricultural technologies has been extensively developed over time, the relevance of the framework rests on its ability to combine almost all types of potential influences; personal, institutional physical and economic. The framework also encompasses adoption and perception in one framework. A slight modification to the framework has however been made to incorporate a feedback mechanism to account for the ex-post

analysis of technology adoption. This aspect is not covered in the original framework which presents the technology adoption process in three interrelated stages: problem awareness (perception), decision stage, and effort stage (adoption) in that order. The ex-post assessment of agricultural technologies is usually conducted by means of impact assessment studies (Maredia et al., 2000; Mercer, 2004; De Janvry et al., 2010), while in information system (IS) technologies, the feedback mechanism is used as the assessment tool (Kim and Malhotra, 2005). The feedback mechanism is an integral aspect, especially in the adoption of voluntary technologies, whereby the resultant attitudes and perceptions are considered to be relatively moderated (Venkatesh and Davis, 2000). Borrowing from the IS field, this feedback mechanism can be used for the assessment of perceptions after the adoption of voluntary technologies in other fields such as the adoption of agri-environmental technologies. It may also follow that the adoption of agro- environmental technologies may have other major benefits for the farmer other than the reduction of water pollution. In such a scenario, adoption of the technology would not be primarily associated with the perception of the water quality problem, something which may only be revealed through the feedback mechanism. Following these arguments and considering that fact that pollution mitigating technologies may have other farm productivity related benefits for the farmers, the conceptual framework with the feedback loop was adopted in this study. A schematic presentation of this framework which indicates the relationship between the dependent and explanatory variables is shown in Fig. 1. 2. Data and methods 2.1. Data The survey data were collected between March and June 2013 among Danish farmers in Funen, Jutland and Zealand regions. An online questionnaire comprising four major sections (farm and farmer demographics, water quality with reference to the WFD requirements, nutrient reduction measures and willingness to pay for constructed wetlands) was distributed to farmers through the regional farmers’ advisory offices. A total of 626 farmers accessed the questionnaire link and responses were received from 368 farmers of which 267 were duly completed. Farmers who dropped out of the survey in the first section of the questionnaire were treated as non-respondents and their information was used in the analysis of unit non-response bias. On the

238

F.G. Gachango et al. / Agricultural Water Management 158 (2015) 235–244

other hand data from farmers who dropped out of the survey in the subsequent sections were used to test for item non-response bias (De Leeuw et al., 2003). The tests indicated non-respondents were mainly older farmers, thereby resulting in an over-representation of young farmers. However, farmers with fully completed and partially completed questionnaires were not statistically different. Although the over-representation of younger farmers could bias the outcome of our analysis, the problem may only be minimal given the current trends of a decrease in the number of older farmers in Denmark (European Commission, 2012). The study therefore utilized the 267 complete cases. Based on four key questions (Appendix A), the analysis was conducted in two stages. Firstly a descriptive analysis of the general respondents’ perception of water quality, attitudes regarding the effects of pollution reduction measures and the preferences for various nutrients mitigation measures implementation strategies was conducted. Secondly, the factors influencing the farmers’ adoption of voluntary nutrient reduction measures and their perception of water quality were analyzed. 2.2. Model specification Based on the adapted conceptual framework, two sets of equations can be specified to represent the perception and the adoption stages. As the model indicates, it is anticipated that the formation of farmers’ perception may have been influenced by their level of adoption of voluntary technologies. Farmers who have voluntarily adopted technologies would therefore be expected to perceive the water quality as being very good and vice versa. This gives the following equations representing the current adoption: ∗ Y1i =



ˇj Xji + ε1i ,

i = 1, 2. . .. . ..n

where ˚ is the normal cumulative distribution function (Green, 2008; Veerbek, 2008). The parameters  and ˇ are estimated by maximum likelihood. The first equation is specified as an ordered probit with the dependent variable (Technology adoption). This variable takes three possible values indicating the number of voluntary technologies already adopted by the farmer such that: 0 = none, 1 = one, 2 = more than one. The farmers were presented with a list of eight technology choices (permanent grass cultivation, biogas, precision farming, organic farming, sediment ponds, natural wetland and constructed wetland). In addition, they had the opportunity to include any other nutrient mitigating technology that was not covered by the provided list. As only a few respondents indicated three or four technologies, these groups were combined with respondents with two technologies thereby reducing the levels to three, as opposed to the original five levels of 0 = none, 1 = one, 2 = two, 3 = three, and 4 = more than three. The second equation is also specified as an ordered probit with the dependent variable “perceived water quality” (Waterquality perception) taking four possible levels in an ordinal format; 1 = low, 2 = moderate, 3 = good and 4 = very good. Due to the low frequency of the “none” and “low” categories in the original dataset, the observations in the two groups are aggregated to “low” category). The choice of independent variables is based on the perception and adoption literature and are broadly classified into four; physical, personal and attitudinal, economic, and institutional factors. A summary of the variables indicating the expected relationship between the explanatory variables and the dependent variables is presented in Table 2. Both models are estimated using the STATA 13 statistical software.

(1) 3. Results and discussion

and perception: ∗ Y2i =



ˇj Xji + Yi∗ i + ε2i ,

i = 1, 2. . .. . ..n

3.1. Descriptive results (2)

∗ and Y ∗ are unobserved. However, what we observe is a where Y1i 2i form of censoring such that:

Y1

⎧ ∗ 0 if Y1i ⎪ ⎨ ⎪ ⎩

1 if 2 if

11 12

≤ 11 ∗ ≤ < Y1i 12 ,

<

(3)

∗ Y1i

and

Y2

⎧ ∗ 0 if Y2i ⎪ ⎪ ⎪ ⎪ ⎨ 1 if 21

≤ 21 , ∗ ≤ , < Y2i 22

∗ ≤ , ⎪ 2 if 22 , < Y2i 23 ⎪ ⎪ ⎪ ⎩ ∗

3 if

23

(4)

< Y2i

Yji ’s

are vectors of the observed explanatory variables, ˇj ’s are the parameters to be estimated corresponding to the X s, ε1i , ε2i and ε3i are error terms (assumed to be normally distributed, N (0,1)), the ’s are the unknown threshold parameters and Yi∗ i are the predicted probabilities obtained from the adoption model estimation. The cumulative probability distribution of each of the Y is therefore given as; Prob(Yi = 0|Xi ) = ˚(−X  i ˇ) Prob(Yi = 1|Xi ) = ˚(i−1 − X  i ˇ) − ˚(−X  i ˇ) Prob(Yi = 2|Xi ) = ˚(i−2 − X  i ˇ) − ˚(i − X  i ˇ) .. . . Prob(yi = J|xi ) = 1 − ˚(j−1 − X  i ˇ)

The average age of the respondents in the study is 51 years and the average farm size is 143 ha. The average distance of the surveyed farms from their nearest recipient water bodies is 14 km with approximately 62% of the farms being within this distance. The majority of the farmers (40%) are in the 45–54 age category indicating a slight difference with the general population where only 33% are in this age category (European Commission, 2014). The average farm size of the sample differs significantly from the average in the general population which is 65 ha. However, the larger farms are critical in the formulation of future policies given the substantial increase in the number of large farms in recent decades (Statistics Denmark, 2014), an aspect that reduces any possible bias. The sample statistics on the different farm types are in line with those of the population with over 90% of the respondents being involved in crop production. Organic and full-time farmers accounts for 9% and 72% of the respondents respectively. With regards to soil types, approximately 50% of the respondents indicate that they have some sandy soils (jb1–jb2) on their farms while 70%, 35% and 25% indicate the presence of clayey-sand (jb3–jb4), sandyclay (jb5–jb6) and heavy-clay (>jb6) soils respectively.1 Medium farm slopes (6–12◦ ) are reported by 56% of the respondents. A summary on the adoption of nutrient reducing technologies indicates that 65% of the respondents are currently employing one or more technologies. Natural wetlands and permanent grass cultivation are the most frequently adopted voluntary pollution reduction technologies at 26% and 25% respectively. Precision

(5) 1 A detailed description of the soil classification in Denmark is available on Greve and Breuning-Madsen (1998).

F.G. Gachango et al. / Agricultural Water Management 158 (2015) 235–244

239

Table 2 Description of variables (N = 267). Dependent variables

Technology Adoption (Model1) Waterquality Perception (Model2)

Variable description

Mean

NPinformation Wetland-funds a

Model1

Model2

0.899

0.850

Perceived water quality

3.750

1.076

0.622

0.485

+

+

0.494 0.259 1.719

0.501 0.439 0.594

+ − +/−

− + +/−

3.393

1.129

+/−

naa

0.813

0.391

+/−

na

51.430 1.532

10.005 0.772

+/− −

+/− +/−

2.411

0.722

+

+/−

2.862

0.606

+

+/−

2.307

0.651

na

+/−

2.582

0.723

na

+/−

2.079 143.440 0.723 0.227

0.944 119.452 0.448 0.338

+/− + − +

na − + na

0.300

0.459

+

+

0.333

0.472

+

+

0.667

0.472

+/−

na

Personal and attitudinal factors Age Farmer’s age (years) Regulation Farmers’ attitude to fines and penalties for non-compliance as a nutrient reduction strategy (3 categories; 1 = non-support, 2 = Indifferent, 3 = Support) Subsidy Farmers’ attitude to subsidy support as a nutrient reduction strategy (3 categories; 1 = non-support, 2 = Indifferent, 3 = Support) Voluntary Farmers’ attitude to voluntary option as a nutrient reduction strategy (3 categories; 1 = non-support, 2 = Indifferent, 3 = Support) Perceived effect on water quality if regulation on fertilizer Nitrogen fertilizer usage is removed (1 = none, 2 = low, 3 = moderate, 4 = high) Animalunit Perceived effect on water quality if regulation on animal unit is removed (1 = none, 2 = low, 3 = moderate, 4 = high)

Institutional factors GES information

Expected sign

Voluntary technologies adoption level

Explanatory variables Exogenous water quality variables/physical factors Distance If farm is located less than 14 km (average sample distance) from the nearest recipient water body (0 = no, 1 = yes) Sand Sandy soil type dummy (0 = no,1 = yes) Heavy-clay Heavy clay soil type dummy (0 = no, 1 = yes) Slope Slope of respondent’s farm: categorical variable (3 levels: 1 = Flat, 2 = Mid-steep, 3 = Steep) Fjord Recipient water body closest to the respondents farm (1 = Hjarbæk, 2 = Kattagat, 3 = Limfjorden, 4 = Mariager 5 = others) Pipedrain If the farm is drained through pipes (0 = no, 1 = yes)

Economic factors Production type Farmsize Farmtype Logturnover

Standard. deviation

Farm types (1 = crop, 2 = crop/cattle, 3 = crop/pig; 4 = others) Size of the farm (ha) Full time farm-type dummy (0 = no, 1 = Yes) Log of turnover averaged over farm-size Subjective farmers’ reception of information on ecological status information (dummy: 0 = no, 1 = yes) Subjective farmers’ reception of information on Nitrogen and Phosphorous reduction measures (dummy: 0 = no, 1 = yes) Farmers’ awareness of the constructed wetlands funding (dummy: 0 = no, 1 = yes)

‘na’ – variable not included in the respective model.

farming and constructed wetlands technologies are adopted by 21% and 4% of the respondents respectively.

3.2. Perception of water quality and effect of regulation removal Based on the five likert point scale, 63% of the respondents perceive the water quality in the water bodies closest to their farms as being above average, whereas a smaller percentage (3%) perceive the quality to be low (Table 3). Farmers’ perceptions regarding the negative impact that removal of existing regulations would have on water quality vary greatly. However, most of the farmers (over 50%) perceive the effect to be low or moderate in all the six scenarios (Table 4). Approximately 20% of the respondents report an expected high negative effect on water quality if regulations on cover crops and winter crops are removed. A descriptive analysis of the farmers’ subjective choice of strategies for the implementation of nutrients reduction measures reveals that the respondents are more supportive of the “voluntary”, “subsidy” and “information dissemination” options with 76%, 55% and 67% respectively. A comparison of respondents’ level

of support for government strategies of implementing the measures, with their current level of adoption of voluntary technologies shows that 62% and 55% of the respondents, who have adopted at least one voluntary technology are in favor of “voluntary” and “information dissemination” strategies respectively. Likewise, 52% of these respondents indicate opposition to the use for fines on non-compliance with the set pollution reduction regulations. The overall summary is presented in Table 5.

3.3. Empirical results The results of the two ordered probit models are presented in Tables 6 and 7. The models Likelihood Ratio Chi-square coefficients (41.40 with 26 degrees of freedom and 51.72 with 24 degrees of freedom) are statistically significant at 0.05 and 0.01 level of probability respectively. This outcome leads us to reject the null hypothesis that all the predictors’ regression coefficients in each model are simultaneously equal to zero. The two models further fulfill the parallel lines assumption at the 0.05 significance level. In each model, a test of whether the levels can be distinguished

240

F.G. Gachango et al. / Agricultural Water Management 158 (2015) 235–244

Table 3 Perception of water quality in the nearest recipient water body. Variable

Water quality

Description

Frequency (%)

Perceived water quality in the nearest water body

None

Low

Moderate

Good

Very good

5.99

3.37

27.34

36.33

26.97

None

Low

Moderate

High

3.00 10.49 7.49 4.49 4.87 5.99

53.56 34.83 57.30 53.56 37.08 37.45

37.08 41.57 32.21 33.33 38.20 48.69

6.37 13.11 3.00 8.61 19.85 7.87

Table 4 Perception of the effect of regulation removal on water quality. Variable

Slurry Phosphorous-feed Nitrogen-fertilizer Pesticides Winter/catch crop Animal unit

Description: regulation removal on

Frequency (%)

Manure/slurry separation, spreading and management Use of high phosphorous quantity in animal feeds Use of nitrogen fertilizers Spraying of pesticides Cultivation of winter crop/catch crops Large animal units

Table 5 Support/non-support for pollution mitigation implementation strategies. Variable

Regulation Zone Voluntary Subsidy Information

Description: level of support for

Frequency (%)

Fines and penalties for non-compliance Specific rules for sensitive areas Use of voluntary programs Use of subsidies Information dissemination

No-support

Indifferent

Support

64.04 22.10 7.49 13.86 7.49

18.73 35.21 16.85 31.09 25.47

17.23 42.70 75.66 55.06 67.04

Table 6 Adoption of voluntary nutrient reduction technologies (Ordered probit model). Variables

Mid-steep slope Age Subsidy -Support Farmsize Wetland funds

Variable description

Medium sloped-farm category Respondent age in years Subsidy support category Farm size in hectares Awareness of constructed wetland funds

1 2 No. Of observations Log likelihood LR 2 (26 d.f.) McKelvey and Zavoina’s R2

Coefficient

0.293* 0.016** 0.417* 0.003*** −0.347** 0.710 2.107 267 −257.140 41.40** 0.180

Standard error

0.160 0.008 0.247 0.001 0.156

Average marginal effects No technology adopted

One technology adopted

Two or more technologies adopted

−0.099 −0.005 −0.143 −0.001 0.114

0.027 0.001 0.045 0.000 −0.024

0.071 0.004 0.098 0.001 −0.090

0.648 0.657

Significance levels; ‘***’ 1%, ‘**’ 5 and ‘*’ 10%.

yields significant differences in the model’s cut points at the 95% confidence interval. The goodness of fit of the models is assessed 2 ). This R2 is more preusing the McKelvey and Zavoina’s R2 (Rmz ferred since it has no bias regardless of the number of categories 2 for the dependent variable. Furthermore, the Rmz values of 0.18 and 0.21 in the adoption and perception models respectively are acceptable within cross-sectional studies (Veall and Zimmermann, 1992). The adoption of voluntary nutrients reduction technologies is significantly explained by the variables on farm slope, farmers’ age, farmers’ attitude to subsidies, farm size and farmers’ awareness of the existence of the constructed wetland funds. Despite the farm size variable being highly significant in explaining the model, the other economic factors included in the model are insignificant. In the perception model, all the physical factors are significant except for the distance variable. None of the personal and attitude variables are found to be significant in determining perceptions of water quality, while economic and institutional factors produce one significant variable each. Overall, the significant variables in

the perception model include soil types, farm slope, full-time farm type and farmers’ subjective reception of information on good ecological status. All the significant variables in both models display the hypothesized direction in their relationship with the dependent variables. Surprisingly, the predicted adoption variable which was expected to significantly influence farmers’ perception on water quality is insignificant in perception model. The results of the estimation coefficients and marginal effects of the significant variables are presented in Tables 6 and 7. Reporting the marginal effects is necessary so as to show the direction of change for all levels of Yi given a change in the Xi ’s. Both the coefficients and marginal effects are used in the discussion of the results although the marginal effects are given more emphasis. 4. Discussion The adoption of one or more voluntary agri-environmental technologies is highly related to farm size. The marginal effects reveal that the probability of having adopted more than one technology

F.G. Gachango et al. / Agricultural Water Management 158 (2015) 235–244

241

Table 7 Perception of water quality in the nearby waterbody (Ordered probit model). Variable

Variable description

Coefficient

Standard error

Average marginal effects

Low perception Sand Heavy-clay Mid-steep slope Steep slope Farmtype GESinformation

Sandy soil dummy (0 = no, 1 = yes) Heavy-clay soil dummy (0 = no, 1 = yes) Medium sloped-farm category Steep sloped farm category Full time farm-type dummy (0 = No, 1 = Yes) Subjective farmers’ reception of information on ecological status information (dummy: 0 = no, 1 = yes)

1 2 3 No. of observations Log likelihood LR 2 (24 d.f.) McKelvey and Zavoina’s R2

Moderate perception

Good perception

Very good perception

−0.248*

0.149

0.037

0.047

−0.010

−0.074

0.289*

0.167

−0.039

−0.056

0.007

0.089

−0.298*

0.175

0.039

0.060

−0.005

−0.094

−0.955*** −0.355*

0.286 0.184

0.185 0.048

0.1500 0.068

−0.092 −0.007

−0.243 −0.110

0.407**

0.176

−0.054

0.006

0.128

0.789 0.304 1.364 267 −320.591 51.72*** 0.207

0.735 0.733 0.734

−0.081

Significance levels: ‘***’ 1%, ‘**’ 5 and ‘*’ 10%.

is higher by 0.1 percentage points for farmers with an average sized farm. This corresponds to other adoption studies (Hodge and Reader, 2010; Ma et al., 2012) where farmers with larger farms are more likely to adopt technologies and schemes that may leave part of the farms out of production. Natural wetlands and permanent grass cultivations which call for setting aside productive land are the most frequently adopted technologies in the current study. The probability of average aged farmers having adopted more than one technology is 0.4 percentage points higher. This finding is however in contrast with other adoption studies (Giovanopoulou et al., 2011; Hodge and Reader, 2010; Ma et al., 2012). These studies however cite high risks and high costs as the main reasons for low technology adoption among older and younger farmers respectively. The widely adopted technologies in this study (permanent grass cultivation and natural wetlands) are however associated with low risks and costs unlike other technologies such as precision farming where uncertainty and high costs are great disincentives for its adoption (Pedersen et al., 2004). The probability of having adopted more than one technology is 7.1 percentage points higher for farmers with mid-steep sloped farms. Surprisingly, farmers with prior information about the constructed wetlands fund are 9.0 percentage points less likely to have adopted more than one technology. This could be an indication of a strong individual motivation for voluntary adoption of environmental technologies. This personal initiative ought to be supported through increased information dissemination of the existing incentives associated with the adoption of agri-environmental technologies (Ryan et al., 2003; Prager and Posthumus, 2010). Additionally, respondents who support subsidies as a government strategy for implementing nutrient reduction measures are more likely to adopt more than one technology by 9.80 percentage points. This finding is in line with Giovanopoulou et al. (2011) who find a lower probability of adopting a nitrate reduction program among farmers who consider subsidy levels to be low. The results of perceived water quality in the descriptive analysis closely agree with the reported quality of water in Danish coastal and inland bathing sites in 2012 with only 3.1% of the

sampled sites being rated poor (European Environmental Agency, 2013). The perception of farmers on the would-be effect on water quality in the event that the regulations are relaxed also corresponds with previous studies on water quality management. Most of the studies show that, generally, farmers do not acknowledge their contribution to water pollution and are not comfortable with the government regulations for controlling pollution from agricultural sources (Macgregor and Warren, 2006; Popp et al., 2007; Buckley, 2012). Macgregor and Warren (2006), observe similar attitudes among farmers operating in nitrate vulnerable zones in Scotland. Popp et al. (2007) find differing perceptions between agricultural and non-agricultural stakeholders while Buckley (2012) identifies skepticism among a group of Irish farmers regarding government regulations. This aspect could also be explained by the overall insignificance of the attitudinal factors in the perception mode. The outcome of the perception model shows that respondents with sandy soil2 on their farms are less likely to perceive high water quality and vice-versa for respondents with farms with heavy-clay portions. These results show that farmers have a clear understanding of soil characteristics and are able to relate this knowledge to the water quality aspects. Clay soils are known to have a good nutrient retention capacity compared to sandy soils, which reduces leaching to ground or surface water (Di and Cameron, 2002; Lehmann and Schroth, 2003). Soils falling under the jb33 classification and above have relatively high clay content (Greve and Breuning-Madsen, 1998). The analysis consequently shows that farmers with mid-steep4 and steeply sloped farms are more likely to perceive low water quality. The overall general significance of the physical factors in the perception model is an indication of farmers’ intrinsic

2 Sandy soil (JB1 and JB2) and heavy.clay (soils above JB6) according to the Danish soil classification. 3 Jb3 soil classification is characterized by a course loamy sand texture. 4 Slope classification in Denmark; flat (<6◦ ), medium steep (6–12◦ ) and steep (>12◦ ).

242

F.G. Gachango et al. / Agricultural Water Management 158 (2015) 235–244

awareness of the role of soil and slope properties in the transportation of agricultural runoff. This awareness therefore plays a key role in forming their attitudes to water quality. Farm characteristics such as slope, soil type and structure have been cited as key factors which contribute to diffuse water pollution (Collins et al., 2007). The coefficient of the dummy indicates that full-time farmers are less likely to perceive the water quality as being very good. The marginal effects further indicate that the probability that full-time farmers perceive water quality as being very good is 11 percentage points lower than among part-time farmers. This implies that the full-time farmers, representing 72% of the respondents, are innately aware of the potential for increased water pollution as a result of their intensive farming activities. The positive and significance coefficient of the variable GESinfo implies that farmers with information on “Good Ecological Status” (GES) of water bodies are more likely to perceive the water quality as being very good. In general the farmers’ awareness and understanding of the concept is expected to guide them in making an informed judgment regarding water quality. Despite the Gesinfo variable significance in the perception model, only 30% of the respondents indicate having prior information on GES. This is an indication of an information gap which could be a result of the centralized system in Denmark (Prager et al., 2011). Similar information gaps have been identified in previous studies and several ways of improving information dissemination have been suggested (Barnes et al., 2009; Blackstock et al., 2010; Barnes et al., 2011). A follow-up question prompting respondents to state their source of information on the GES requirements shows municipalities as the key information point with 69% of the respondents having received information through their local authorities. Approximately 44% of the respondents indicate having received the information from agricultural advisors, while the media and environmental agency are rated as sources of information by 47% and 38% of the respondents respectively. A small group of farmers, on the other hand report having received the information from other farmers, research and academic institutions and from the Internet. The predicted variable on the adoption of voluntary nutrient reduction technologies shows an insignificant relationship with perception. This outcome may have resulted from the fact that some of the technologies adopted by the farmers have multiple benefits, thus yielding different utilities to the adopters. Practicing precision farming is one such example where farmers may employ site specific fertilizer and pesticide applications, thereby reducing the variable production costs in addition to reducing excess nutrients that may end up in farm discharge. Some previous studies (Innes and Sam, 2008; Barnes et al., 2011) have found some farmers and firms adopting pollution mitigating technologies in parallel to the regulatory measures.

to be above average. Consequently, the farmers feel that changes that would lead to the removal of the regulatory measures would have little or no effect in terms of reducing the water quality. This perception may however be misinformed, since majority of the farmers have not had access to full information on the underlying stipulations and requirements of “good ecological status” under the Water Framework Directive. The study also finds that, in general, farmers have a negative attitude towards regulatory nutrient reduction measures despite the fact that a relatively large number of these farmers have adopted voluntary environmental measures. The adoption of voluntary technologies is influenced by farm size and farmers’ attitudes to subsidies. Additionally, other factors such as farm type, soil types and farm slope play a key role in influencing farmers’ perceptions and need to be fully considered in the implementation design of the more targeted pollution reduction measures. To bridge the identified information gap, a collaborative effort between policy makers and the various stakeholders to decentralize information should be pursued. This will ensure a smooth and efficient flow of information, thereby improving the implementation of measures and increasing the likelihood of achieving the water quality target. Whereas the regulatory measures remain in place, designing tailor-made incentives based on farm structure and physical characteristics would greatly improve the pace of adoption of the technologies thereby reducing water pollution substantially. Acknowledgements This research has been conducted under two Danish Research Projects; SupremeTech – a Strategic Research Project, Supported by the Innovation Fund, and iDRÆN – a Research Project funded by GUDP, Danish Ministry of Food, Agriculture and Fisheries. We are grateful to Flemming Gertz of SEGES for his valuable contribution in coordinating the survey process. Appendix A. Questionnaire extract Question

Choices

1.

How would you rate the quality of water in this water body (referring to the water body mentioned in the previous question)

2.

If the regulations related to the farm activities listed below were removed, what level of effect (negative impact) do you think this would have on the quality of water in the recipient water body closest to your farm? • Slurry/manure separation and application • Use of high phosphorous animal feeds • Use of nitrogen fertilizers • Spraying of pesticides • Cultivation of winter crop/catch crops • Keeping large animal units To what extent would you support (agree with) the use of the different implementation methods of nutrients reduction measures listed in the table below? • Use of fines for non-compliance • Zoning (different zones are assigned different measures) • Voluntary programs • Payment of subsidies • Information dissemination

 Low  Moderate  Good  Very good  None/do not know  Low  Moderate  High  None/do not know

5. Conclusions and policy implications Understanding farmers’ decisions on the adoption of voluntary nutrients pollution mitigating technologies and their inherent attitudes and perceptions regarding water quality and the existing regulatory measures are key starting points to consider when designing and providing the options for measures that reduce pollution from agricultural fields. This study has analyzed these factors in order to identify the critical aspects to consider in the future design of the constructed wetland measure, and its implementation among Danish farmers to ensure a high adoption rate of the measure, thereby achieving policy success. The study finds that the majority of the farmers perceive the quality of water in the fjords and lakes draining from their farms

3.

 Not support  Indifferent  Support

F.G. Gachango et al. / Agricultural Water Management 158 (2015) 235–244

4

Question

Choices

What measures do you practice in your farm to help in the reduction of water pollution? (tick all measures that you practice))

 Grassland cultivation  Biogas production  Organic farming  Precision farming  Sedimentation ponds  Natural wetland  constructed wetlands  Other (specify)

References Ahnström, J., Höckert, J., Bergeå, H.L., Francis, C.A., Skelton, P., Hallgren, L., 2009. Farmers and nature conservation: What is known about attitudes, context factors and actions affecting conservation? Renewable Agriculture and Food Systems 24 (01), 38–47. Ajzen, I., 1975. From Intentions to Actions: A Theory of Planned Behavior. Springer Berlin, Heidelberg, pp. 11–39. Ajzen, I., Fishbein, M., 1975. Belief, Attitude, Intention and Behavior: An Introduction to Theory and Research. Addison-Wesley, Reading, MA. Atkins, J.P., Burdon, D., 2006. An initial economic evaluation of water quality improvements in the Randers Fjord, Denmark. Marine Pollution Bulletin 53 (1), 195–204. Balticdeal, 2011. http://www.balticdeal.eu/measures/ (accessed 30.03.14). Barnes, A.P., Willock, J., Hall, C., Toma, L., 2009. Farmer perspectives and practices regarding water pollution control programmes in Scotland. Agricultural Water Management 96 (12), 1715–1722. Barnes, A.P., Willock, J., Toma, L., Hall, C., 2011. Utilising a farmer typology to understand farmer behaviour towards water quality management: Nitrate Vulnerable Zones in Scotland. Journal of Environmental Planning and Management 54 (4), 477–494. Beedell, J., Rehman, T., 2000. Using social-psychology models to understand farmers’ conservation behaviour. Journal of Rural Studies 16 (1), 117–127. Blackstock, K.L., Ingram, J., Burton, R., Brown, K.M., Slee, B., 2010. Understanding and influencing behaviour change by farmers to improve water quality. Science of the Total Environment 408 (23), 5631–5638. Bratt, A., 2002. Farmers’ choices: management practices to reduce nutrient leakage within a swedish catchment. Journal of Environmental Planning and Management 45 (5), 673–689. Buckley, C., 2012. Implementation of the EU Nitrates Directive in the Republic of Ireland—a view from the farm. Ecological Economics 78, 29–36. Danish Ministry of Environment, 2009. Agreement for Green Growth. Environmental Protection Agency, Denmark. De Leeuw, E.D., Hox, J.J., Huisman, M., 2003. Prevention and treatment of item nonresponse. Journal of Official Statistics 19 (2), 153–176. De Janvry, A., Dustan, A., Sadoulet, E., 2010. In: Recent Advances in Impact Analysis Methods for Ex-Post Impact Assessments of Agricultural Technology: Options for the CGIAR. University of California at Berkeley, http://impact. cgiar. org/ meetings-and-events Di, H.J., Cameron, K.C., 2002. Nitrate leaching in temperate agroecosystems: sources, factors and mitigating strategies. Nutrient Cycling in Agroecosystems 64 (3), 237–256. Christensen, T., Pedersen, A.B., Nielsen, H.O., Mørkbak, M.R., Hasler, B., Denver, S., 2011. Determinants of farmers’ willingness to participate in subsidy schemes for pesticide-free buffer zones—a choice experiment study. Ecological Economics 70 (8), 1558–1564. ˝ Collins, A.L., Stromqvist, J., Davison, P.S., Lord, E.I., 2007. Appraisal of phosphorus and sediment transfer in three pilot areas identified for the catchment sensitive farming initiative in England: application of the prototype PSYCHIC model. Soil Use and Management 23 (s1), 117–132. Ervin, C.A., Ervin, D.E., 1982. Factors affecting the use of soil conservation practices: hypotheses, evidence, and policy implications. Land Economics 58 (3), 277–292. European Commission, 2012. Eurostat: Category – Agriculture. http://ec.europa. eu/eurostat/statistics-explained/index.php/Category:Agriculture (accessed 20.01.14). European Commission, 2014. Member State Factsheet Denmark: Agriculture and Rural Development. http://ec.europa.eu/agriculture/statistics/ (accessed 15.07.14). European Environmental Agency, 2013. Environmental Indicator Report 2013: Natural Resources and Human Well-being in a Green Economy. http://www.eea. europa.eu/publications (accessed 30.01.14). Giovanopoulou, E., Nastis, S.A., Papanagiotou, E., 2011. Modeling farmer participation in agri-environmental nitrate pollution reducing schemes. Ecological Economics 70 (11), 2175–2180. Greve, M.H., Breuning-Madsen, H., 1998. Soil Mapping in Denmark. European Soil Bureau. Research – Report No. 6, pp. 49–57. Greene, W.H., 2008. Econometrics Analysis, 6th edition. Pearson Pentice Hall, Upper Saddle River, New Jersey. Hodge, I., Reader, M., 2010. The introduction of Entry Level Stewardship in England: extension or dilution in agri-environment policy? Land Use Policy 27 (2), 270–282.

243

Hoffmann, C.C., Kjaergaard, C., Levesen, B., Cochran, M., 2011. Comparing two surface flow wetlands for removal of nutrients in agricultural drainage water. In: Science for the Environment-Environment for Society. Aarhus, Denmark. Hu, Z., Morton, L.W., 2011. US midwestern residents perceptions of water quality. Water 3 (1), 217–234. Innes, R., Sam, A.G., 2008. Voluntary pollution reductions and the enforcement of environmental law: an empirical study of the 33/50 program. Journal of Law and Economics 51 (2), 271–296. Jacobsen, B.H., Laugesen, F.M., Dubgaard, A., Bojesen, M., 2013. Biogasproduktion i Danmark–Vurderinger af drifts-og samfundsøkonomi. Institut for Fødevare-og Ressourceøkonomi, Københavns Universitet. IFRO Rapport, no. 220. Just, R.E., Zilberman, D., 1983. Stochastic structure, farm size and technology adoption in developing agriculture. Oxford Economic Papers, 307–328. Kaplowitz, M.D., Witter, S.G., 2008. Agricultural and residential stakeholder input for watershed management in a mid-Michigan watershed. Landscape and Urban Planning 84 (1), 20–27. Kabii, T., Horwitz, P., 2006. A review of landholder motivations and determinants for participation in conservation covenanting programmes. Environmental Conservation 33 (01), 11–20. Kim, S.S., Malhotra, N.K., 2005. A longitudinal model of continued IS use: an integrative view of four mechanisms underlying post adoption phenomena. Management Science 51 (5), 741–755. Kjærgaard, C., Canga, E., Heckrath, G.J., Iversen, B.V., Hoffmann, C.C., Druenhahl, J.B., Nielsen, P., Gertz, F., Lyngsie, G., Strobel, B., Borggaard, O., Hansen, H.C.B., 2012. Drainage filters and constructed wetlands to mitigate site specific nutrient losses. In: Proceedings from the 7th SWS 2012, European Chapter Meeting. Kjaergaard, C., Hoffmann, C.C., 2013. Constructed Wetlands for Targeted Reduction of Nutrients in Drainage Water, No. 61202, 33 pp. Knowler, D., Bradshaw, B., 2007. Farmers’ adoption of conservation agriculture: a review and synthesis of recent research. Food Policy 32 (1), 25–48. Lehmann, J., Schroth, G., 2003. Nutrient Leaching. Trees, Crops, and Soil Fertility: Concepts and Research Methods. CAB International, Wallingford, UK, pp. 151–166. Lichtenberg, E., Lessley, B.V., 1992. Water quality, cost-sharing, and technical assistance: perceptions of Maryland farmers. Journal of Soil and Water Conservation 47 (3), 260–264. Ma, S., Swinton, S.M., Lupi, F., Jolejole-Foreman, C., 2012. Farmers’ willingness to participate in payment-for-environmental-services programmes. Journal of Agricultural Economics 63 (3), 604–626. Macgregor, C.J., Warren, C.R., 2006. Adopting sustainable farm management practices within a Nitrate Vulnerable Zone in Scotland: the view from the farm. Agriculture, Ecosystems and Environment 113 (1), 108–119. Maredia, M., Byerlee, D. and Anderson, J.R., 2000. Ex Post Evaluation of Economic Impacts of Agricultural Research Programs: A Tour of Good Practice. Mbaga-Semgalawe, Z., Folmer, H., 2000. Household adoption behaviour of improved soil conservation: the case of the North Pare and West Usambara Mountains of Tanzania. Land Use Policy 17 (4), 321–336. Mercer, D.E., 2004. Adoption of agroforestry innovations in the tropics: a review. Agroforestry Systems 61 (1–3), 311–328. Ministry of Food, Agriculture and Fisheries, 2008. Report on Climate and Agriculture. http://fvm.dk/foedevarer/indsatsomraader/klima/rapport-om-landbrugog-klima/ (accessed 30.05.13). Ministry of Food, Agriculture and Fisheries, 2012. The Danish Rural Development Programme 2007–2013. http://naturerhverv.dk/tvaergaaende/eu-arbejdet/ landdistriktsprogrammet-2007-2013/ (accessed 30.05.13). Minivådområder, 2012. Minivådområder udviklingsprojekt. http://www. minivaadomraader.dk/ (accessed 10.04.14). Mojo, D., Todo, Y., Matous, P., 2010. Perception of farmers and agricultural professionals on changes in productivity and water resources in Ethiopia. World Academy of Science, Engineering and Technology 1 (66), 860–868. Morton, L.W., 2007. A different view: Mid-West farmers’ perceptions of water quality. In: Annual Meeting of the Rural Sociological Society, vol. 3, Marriott Santa Clara, Santa Clara, California. Nguyen, V.T., Montaz, S., Zimmerman, K., 2006. Water pollution concerns in shrimp farming in vietnam: a case study of Can Gio, Ho Chi Minh City. International Journal of Environmental, Cultural, Economic and Social Sustainability 3 (2), 129–138. Pedersen, S.M., Fountas, S., Blackmore, B., Gylling, M., Pedersen, J.L., 2004. Adoption and perspectives of precision farming in Denmark. Acta Agriculturae Scandinavica, Section B-Soil & Plant Science 54, 2–8. Perez-Espejo, R., Ibarra, A.A., Escobedo-Sagaz, J.L., 2011. Agriculture and water pollution: farmers’ perceptions in Central Mexico. Water Resources Development 27 (01), 263–273. Popp, J., Rodriguez, G., Gbur, E., Pennington, J., 2007. The role of stakeholders’ perceptions in addressing water quality disputes in an embattled watershed. Journal of Environmental Monitoring and Restoration 3, 225–263. Prager, K., Posthumus, H., 2010. Socio-economic factors influencing farmers’ adoption of soil conservation practices in Europe. Human dimensions of Soil and Water Conservation: A Global Perspective. Nova Science Publishers, Inc, New York, pp. 203–223. Prager, K., Schuler, J., Helming, K., Zander, P., Ratinger, T., Hagedorn, K., 2011. Soil degradation, farming practices, institutions and policy responses: an analytical framework. Land Degradation and Development 22 (1), 32–46. Ryan, R.L., Erickson, D.L., De Young, R., 2003. Farmers’ motivations for adopting conservation practices along riparian zones in a mid-western agricultural watershed. Journal of Environmental Planning and Management 46 (1), 19–37.

244

F.G. Gachango et al. / Agricultural Water Management 158 (2015) 235–244

Sang, N., 2008. Informing common pool resource problems: a survey of preference for catchment management strategies amongst farmers and the general public in the Ythan river catchment. Journal of Environmental Management 88 (4), 1161–1174. Savage, J.A., Ribaudo, M.O., 2013. Impact of environmental policies on the adoption of manure management practices in the Chesapeake Bay watershed. Journal of Environmental Management 129, 143–148. Smith, C.M., Peterson, J.M., Leatherman, J.C., 2007a. Attitudes of Great Plains producers about best management practices, conservation programs, and water quality. Journal of Soil and Water Conservation 62 (5), 97A–103A. Smith, R.J.N., Glegg, G.A., Parkinson, R., Richards, J.P., 2007b. Evaluating the implementation of the Nitrates Directive in Denmark and England using an actor-orientated approach. European Environment 17 (2), 124–144.

Statistics Denmark, 2014. http://www.dst.dk/en/Statistik/emner/landbruggartneri-og-skovbrug.aspx (accessed 15.09.14). Veall, M.R., Zimmermann, K.F., 1992. Pseudo-R 2’s in the ordinal probit model*. Journal of Mathematical Sociology 16, 333–342. Veerbek, M., 2008. A Guide to Modern Econometrics, 3rd edition. John Wiley and Sons Ltd., The Atrium South Gate, West Sussex. Venkatesh, V., Davis, F.D., 2000. A theoretical extension of the technology acceptance model: four longitudinal field studies. Management Science 46 (2), 186– 204. Wauters, E., Bielders, C., Poesen, J., Govers, G., Mathijs, E., 2010. Adoption of soil conservation practices in Belgium: an examination of the theory of planned behaviour in the agri-environmental domain. Land Use Policy 27 (1), 86–94.