Peat Land Oil Palm Farmers’ Direct and Indirect Beneﬁts from Good Agriculture Practices

: With economically unsustainable metroxylon sagu (sago palms) found in peat lands, small scale farmers are gradually converting their land to oil palm cultivation. Good agriculture practices (GAP) were inculcated to peat land farmers to ensure that the environmental ecosystem is conserved and oil palm productivity is enhanced, along with the farmer’s well-being. The present study examined the effect of GAP on farm performance and the perceived economic well-being of the peat land oil palm farmers. We interviewed randomly selected farmers with assistance from a locally trained native enumerator to complete the survey questionnaire. We conducted partial least square structural equation modeling (PLS-SEM) to incorporate direct and indirect beneﬁts on farmers’ economic well-being to estimate the signiﬁcance of GAP. The empirical results show that GAP have direct positive effects on farm performance. Such practices lead to signiﬁcant positive impacts on the economic well-being of peat land oil palm farmers. This solid evidence makes it much easier for small-scale farmers to convert from conventional farming to environmentally friendly farming, and ensures safe and healthy oil palm cultivation.


Introduction
Peat lands are significant landscapes for the environment, economy, and public health. Peat lands cover an estimated area of 400 million hectares of the Earth's land surface, whereas tropical peat lands cover only 30-40 million hectares [1]. The soil found in peat lands is dominated by organic soil materials, with more than 50% in the upper layer of the soil [2,3]. The major typical problems with peat soil, especially deep peat, lie in its physical and chemical properties. Peat land ecosystems have various functions, especially for storm water retention, flood protection, water quality enhancement, freshwater fishery, food chain support, feeding grounds for marine fish, biodiversity, carbon storage, and climate mitigation [4][5][6]. Loss of biodiversity [7,8] and the contribution to global climate change through carbon emissions are the main concerns about peat land conversion [9][10][11][12][13]. Therefore, changes in peat land use and land cover affect the ability of landscapes to continue providing the high-quality ecosystems required for human health and wellbeing [14,15]. Some peat swamp areas have proven suitable for oil palm cultivation; whereas, some areas need to be conserved because of their unsuitability for agricultural

Related Theory
Two main theories were used to discuss the adoption of GAP; namely, technology dissemination theory [50] and adult learning theory [51,52]. According to Rogers [50], a farmer decides to make full use of new technologies in farming practices for productivity enhancement and solving their problems [53,54]. With transformative learning theory [55], we expected that peat land farmers could gain new knowledge and practices to resolve challenges and sustain their future economic well-being. The benefits of good agricultural skills, attitude, knowledge, and practices [13,43,49,56,57], such as high productivity, added value, and resistant crops [54,[58][59][60], were communicated to farmers. Transformative learning leads farmers to use past experiences as a guide for future actions, to be reflective, open to new perspectives, less defensive, and receptive to new ideas [55,61] for increasing productivity. Davis et al. [62] and Rustam [63] and Leitgeb et al. [64] suggested that farmers can learn in groups and collaborate on discovering and solving their agriculture problems.

Benefits in Farm Performance
A systematic literature review shows that environmental sustainability practices in oil palm cultivation enriched pollinator populations, which enhanced pollination [65]. Bee pollination improves the yield of most crop species, including oil palm crops [65,66]. Adequate weed management will protect the soil from erosion and provide a habitat for natural pests, while interacting with water and nutrient cycles [67]. An optimal yield is possible when oil palm trees are healthy [68]. Good drainage, which is needed to sustain oil palms in peat land, protects against the subsidence of soil, the potential deepening of future flood risks, and reduced oil palm growth [19]. Farm drainage systems, lanes, and farm roads should be well constructed and maintained. Groundwater should also be maintained to minimize peat subsidence [11,69,70], peat oxidation, and CO 2 emissions, as well as to drain out excess water for superior palm growth and production of fresh fruit bunches (FFB) [11,20,71]. The water level within drainage flows should be regulated frequently. Healthy oil palm trees are those not infected with parasites, pests, and diseases, such as fungus and Ganoderma; due to GAP in the selection of plant material, well-planned fertilization, pest control, and irrigation. Moreover, soil fertility needs to maintained to increased crop yields [26]. Previous research has shown that practicing good agricultural holistically increased agricultural yields. As a result, and as displayed in Figure 1, GAP are hypothesized to have a direct and positive effect on farm performance. agricultural extension programs that were aimed at increasing farmers' knowledge of environmental consciousness enhanced their productivity. Previous literature has demonstrated that GAP improved farm performance in terms of yield, sufficient nutrients, and the absence of diseases, pests, and parasites. We concluded that oil palm productivity has a significant and positive effect on per capita income. Farmers' economic well-being is enhanced by better farm performance, as shown in Figure 1, which generates sufficient income for their daily needs. As illustrated in Figure 1, good agricultural practices are hypothesized to have indirect and direct positive impacts on farmers' economic well-being. Thus, this study made the following hypotheses.

Research Site, Population, and Sample Size
Dalat in the Mukah division was the research site of this study (in Appendix A). It is on the northwest coast of Sarawak in East Malaysia and covers an area of approximately 0.7 million hectares. This division is inhabited by 122,300 people, predominantly by the Melanau ethnic group, who comprise approximately 60% of the total Mukah population [86]. Only 5.6% of the total land of Sarawak is composed of low-lying peat. Since the 1880s, sago palm (Metroxylon sagu) has been the major commercial crop in this area and other parts of Sarawak [87]. Sago palm is classified as a high-energy-yielding crop, because it

Benefits for Farmers' Economic Well-Being
Oil palm plantations do not only provide farmers with permanent income but also job opportunities in rural areas. Planting oil palm in peatland areas can generate direct income for farmers if managed properly [22]. Ali and Sharif [72], and Waddington, White, and Anderson [68] found that the transfer of technology of environmentally friendly cul-tivation has affected crop performance and improved the well-being of farmers. With improved oil palm productivity, the quality of FFB leads to an increase in the income of farmers [25,31,73]. Empirical evidence from Indonesia shows that oil palm farmers with GAP enjoy higher yields than those who performed poorly [74,75]. Thus, high crop performance is associated with the high economic well-being of farmers [25,26]. If properly managed and enforced, sustainable oil palm cultivation can boost productivity [12,19,76,77]. Tilman et al. [78]; Feder, Murgai, and Quizon [79]; Woittiez et al. [76]; Awang Besar et al. [29]; and Nong et al. [80] suggested the transfer of technology to farmers through sustainable farm management practices, including reducing the productivity gap. A GAP is a set of norms and procedures developed to produce agricultural products in a sustainable manner [57,67,81]. Paramananthan [82] stated that oil palm productivity is also induced by environmental protection. Mariyono [83], and Yamazaki and Resosudarmo [84] showed that the use of integrated pest management leads to reliable productivity. Similarly, Mancini et al. [85], Obidzinski [21], and Saadun et al. [46] proved that agricultural extension programs that were aimed at increasing farmers' knowledge of environmental consciousness enhanced their productivity. Previous literature has demonstrated that GAP improved farm performance in terms of yield, sufficient nutrients, and the absence of diseases, pests, and parasites. We concluded that oil palm productivity has a significant and positive effect on per capita income. Farmers' economic well-being is enhanced by better farm performance, as shown in Figure 1, which generates sufficient income for their daily needs. As illustrated in Figure 1, good agricultural practices are hypothesized to have indirect and direct positive impacts on farmers' economic well-being. Thus, this study made the following hypotheses.

Research Site, Population, and Sample Size
Dalat in the Mukah division was the research site of this study (in Appendix A). It is on the northwest coast of Sarawak in East Malaysia and covers an area of approximately 0.7 million hectares. This division is inhabited by 122,300 people, predominantly by the Melanau ethnic group, who comprise approximately 60% of the total Mukah population [86]. Only 5.6% of the total land of Sarawak is composed of low-lying peat. Since the 1880s, sago palm (Metroxylon sagu) has been the major commercial crop in this area and other parts of Sarawak [87]. Sago palm is classified as a high-energy-yielding crop, because it needs high infrastructure costs due to remoteness and caters mainly to a limited domestic demand [88]. Melling [33] found that the virgin peat swamp in this area is unsuitable for sago palm and causes stunted growth and a high mortality rate. Growing sago palm is also an economically unsustainable industry [89], because its maturity period is almost 10 years [90]. Such a long period can worsen poverty among farmers. Thus, farmers gradually convert their land to oil palm cultivation, which has become a source of permanent income and employment in this rural area. By 2016, approximately 43% of all farmers had planted oil palm on peat land area. The oil palms planted in peat soil area by farmers are mostly concentrated in the Matu (87.0%) and Daro (100%) districts in the Dalat area. On the basis of the calculations of Yamane (cited from Israel [91]), the minimum target sample size in this study was 200 peat land oil palm farmers in four villages. We randomly selected the independent farmers who planted oil palm in this peat land area. Based on the sample size, we generated random numbers and chose a peat land farmer who was registered with MPOB. Roscoe [92] stated that the applicable sample size for research is between 30 to 500 farmers, whereas Kwong [93] suggested a minimum sample size of 59 farmers, with three arrows pointing at latent variables. Hoyle and Gottfredson [94] suggested that a sample size of 100 to 200 farmers is a good start in path analysis. We conducted face-to-face interviews with independent oil palm farmers and were assisted by a locally trained enumerator [95] to complete the tested survey questionnaire. Fieldwork was conducted in August-October 2019. Finally, we obtained 78.5% (157) completed survey questionnaires that were reliable for further analysis.

Instrument Design
The economic well-being of farmers involved in the conversion of peat land to oil palm plantations is investigated in this study. The economic well-being of peat land oil palm farmers is determined by GAP and mediated by farm performance. The most important measure in agronomic performance is oil palm yield, which depends on GAP. We measured GAP in peat land cultivation, primarily adopted from the Food and Agriculture Organization Good Agricultural Practices [24,46]. We combined the indicators on the basis of a diagnostic smallholder constructed by Woittiez et al. [96]. Then, we classified GAP systematically into six dimensions; namely, fertilization techniques, harvest and pruning techniques, drainage and transportation, parasite control and planting system, soil and water quality, and farm and financial management (in Table 1), as pursued by Choo and Abu-Bakar [2], and Woittiez et al. [76]. Agricultural yield and income generated from agricultural production are often key indicators of a farms' well-being and economic performance [12,25,68,72,74,[97][98][99]. Oil palm FFB (ton/ha) are the immediate products of oil palm farmers. These are then processed into crude palm oil and crude palm kernel oil by millers. An optimal yield is possible when oil palm trees are healthy. A healthy oil palm tree does not show any signs of major nutrient deficiency, such as yellow leaves, purple stems, orange spots, dry lower leaf blades, and wrinkled leaves. A healthy oil palm tree is also not infected with parasites, pests, or diseases, such as fungus and Ganoderma. Self-perception questions were used in this section, such as "Oil palms are free from pests and parasites". We used three-point Likert scales of possible answers to measure farm performance (1 = never, 2 = sometimes, and 3 = always). Table 1 shows farmers' economic well-being measured in subjective measurements. The subjective economic well-being consists of farmers' perceptions of their earnings, finances, and wealth derived from oil palm cultivation. The earnings of oil palm farmers are highly dependent on yield and the price of oil palm fresh bunches. Income generated from oil palm sales [25,68,73,98,100] reflects the financial benefits enjoyed by farmers. We selected and adopted subjective well-being items on income sufficiency/adequacy and financial stability from previously established measurements [101][102][103]. The item responses in the economic well-being scale ranged from 1 = disagree to 3 = agree [104]. The farmers selected the Likert scale that best described them, such as "income generated from oil palm cultivation is sufficient for a family's daily meals".

Latent Variable Validity and Reliability
With 35 farmers participating in the pilot test, the study revealed that the Cronbach's alpha reliability coefficient (α) for economic well-being was 0.93, farm performance was 0.70, and the GAP was 0.90. The structured survey questionnaires were also based on actual data gathered for convergent validity (in Table 1), using partial least squares structural equation modeling analysis (PLS-SEM). This study used SmartPLS as a statistical analysis software to run PLS-SEM and evaluate the measurement quality and structural model. Hair, Ringle, and Sarstedt [105], and Kwong [93] proposed that composite reliability should be more than 0.70, whereas 0.60 is acceptable. By using PLS-SEM analysis [106,107], we identified the composite reliability for GAP, farm performance, and economic well-being of the oil palm farmers. Most of the outer loadings of the items were more than 0.70, which was considered good and reliable [105]. Only one item was less than 0.70 but higher than 0.40; nevertheless, it was acceptable [106]. A factor loading, which was greater than 0.50, was used as a criterion to select a statement in a factor [105]. Chin [108] pointed out that loads of at least 0.50 were acceptable. Average variance extracted (AVE) should be 0.50 or more [109]. Table 1 shows that the composite reliability and AVE were greater than 0.70 and 0.50, respectively, exceeded the minimum acceptable values, and proved good internal consistency for each latent construct [110].
Two methods can be used to determine discriminant validity [107,[111][112][113]. In the first method, the AVE is examined. The AVE value must be greater than 0.50 [114]. In this study (Table 1), AVE was calculated to assess the discriminant validity of the eight latent constructs, which ranged from 0.57 to 0.82. These data showed that all values of AVE were above 0.50. Moreover, the square root of the AVE between each pair of factors was higher than the correlation projected between factors, thereby ratifying the discriminate validity. Alternatively, the square roots of AVE were compared with those of the other constructs below the diagonal in Table 2. These statistics suggest that each construct is stronger in its own measurement than in the measurement of another construct [106]. The statistics suggest that the elements of our measurements are reliable, internally consistent, and have discriminant validity. Table 2 shows the discriminant validity of the constructs. The comparison of cross-loadings in Table 3 shows that the loadings of an indicator are higher than the other loadings for its construct in the same column and row. Furthermore, the results indicate that discriminant validity exists among all the constructs on the basis of the loadings depicted in Table 2.

Model Verification
The PLS-SEM was used as a statistical technique for estimating the cause-and-effect relationship models using latent variables [97,107]. A combination of multiple regression and PLS-SEM is ideal for data analysis to assess the structural effect GAP of peat land oil palm plantations on the economic well-being of farmers. We used the Smart PLS statistical software to take advantage of cross-sectional data [93,115] to identify the direct and indirect effects of GAP on the oil palm farmers' economic well-being. Smart PLS, which was developed in Java, is an appropriate software to the analyze data, as suggested by Hair et al. [106], Hair Jr et al. [115], and Hair, Hult, and Ringle [107].

Oil Palm Smallholder Profile
We predominantly used percentages to describe peat land farmers' background and farm characteristics. Almost half of independent oil palm farmers have their own land, followed by 24.2% with mortgaged land, 18.3% with heritage land, and 7.8% with other types of land ownership (in Appendix B). Half of peat land oil palm farmers do not have any formal education, 19.7% only reached primary school, and 15.3% reached lower secondary school. Only 7.0% of farmers are high school graduates, 1.3% are diploma holders, and 1.3% are degree holders. Only 14.0% of the peat land oil palm farmers were less than 39 years old, more than half (55.5%) were 40-59 years old, and the rest (30.6%) were more than 61 years old. With limited oil palm land and small-scale production, 66.7% of farmers earn a monthly cash income between MYR 501 to MYR 1500, and 13.0% earn MYR 1501 to MYR 2500. Only 15.2% of farmers earn less than a MYR 500 of monthly cash income generated from oil palm cultivation. However, peat land oil palm farmers do not integrate crops or livestock into their farms (M = 1.834; SD = 0.6186) to diversify their income. On average, oil palm farmers in this study area have 3.64 hectares. With the latest agricultural innovations and strengthened indigenous knowledge, local farmers are able to optimize their sustainable cultivation and management practices [53,68,[116][117][118]. Farmers need to be trained further to solve their cultivation challenges in a sustainable manner [53,68,79,119,120].

Effects on Farm Performance and Farmers' Economic Well-Being
Smart PLS 3.0 was used to examine the structural model and hypothesis [107]. The path estimates and t-statistics were calculated for the hypothetical relationships by using an algorithm and bootstrapping technique with a resampling of 5000. The structural model (Figure 2) shows the relation between one variable and another variable with beta (β) and R-squared (R 2 ) values. The results showed that the R 2 for farm performance was 0.05, and the R 2 for farmers' economic well-being was 0.13. The R 2 value of economic wellbeing of farmers could be explained or influenced by 13.0% of the independent variables (GAP and farm performance), and the rest were influenced by other factors outside this model. In general, a combination of adequate rainfall, sunshine, ideal soil conditions, good weed management, fertilization, irrigation, pre-harvest and post-harvesting, and pest and disease control would result in optimal growth and yield of oil palms [19,65,73]. Oil palm plantations in these areas have soft weeds and are free of shrubs, which led to positive effects. The results showed that GAP (β = 0.23, p < 0.01) had a significantly positive and direct effect on farm performance, as predicted in H1. We found that farm performance has a positive and direct significant effect on peat land farmers' economic well-being. The results are consistent with H2, that farm performance (β = 0.26, p < 0.01) was positively related to farmers' economic well-being. However, the integration of crop and livestock (in Appendix B) is typically practiced less often by peat land farmers, which could increase weed control and the productivity of oil palms [121,122]. Similarly, crop and livestock integration in oil palm plantations generates additional income and employment [99,100,123].

Direct and Indirect Effects on Farmers' Economic Well-Being
On the basis of the above literature and framework, we tested the hypothesis (H3) that GAP have a positive and direct significant effect on farmers' economic well-being, and (H4) that GAP have a positive and indirect significant effect on farmers' economic well-being. Figure 2 shows that GAP (β = 0.18, p < 0.01) is positively and directly related to farmers' economic well-being. Furthermore, GAP is significantly indirect related to the economic well-being (β = 0.06, p < 0.05) of peat land oil palm farmers and is mediated by crop performance. This result strengthened the findings of Cole and Fernando [124], and Mkanthama et al. [125], that eco-healthy agricultural practices, such as nutrient management, seed selection, and weed management, improved productivity. In line with Qaim et al. [12], Luke et. al. [40], Uning et al. [42], Santika et al. [126], and Heylen et al. [127] we confirmed that GAP are associated with diminished poverty among farmers. Table 4 shows that the total effects of GAP on farm performance and the economic well-being of peat land oil palm smallholders can be classified as complementary mediation. This finding implied that all paths in the model have a strong effect on farmer's economic well-being, as shown in Figure 2 and Table 4. This finding confirmed previous works [26,66,68,76], with consistency in the results. Healthy oil palm trees lead to high productivity and high-quality FFB. Thus, optimal yield will lead to a positive impact on the income of farmers [31,74,75]. Similarly, farm performance has a significant impact on the well-being of peat land oil palm farmers. Peat land oil palm farmers' economic well-being is also significantly and directly affected by GAP. GAP in peat land oil palm plantations has indirect impacts on economic well-being and is mediated by crop performance.

Conclusions
We analyzed the GAP among farmers of peat land oil palm, and it effects on farm performance and economic well-being. We conducted a path analysis of direct and indirect effects on farmers' economic well-being, which is key to understanding the effectiveness of GAP in peat land oil palm cultivation. The empirical results showed that GAP have direct positive effects on farm performance and positive total effects on economic well-being.
The GAP of small-scale peat land farmers needs to be further intensified. Irrespective of the planting and maintenance process; harvesting and pruning techniques, drainage and transport, and parasite control and planting systems, although functional, need to be enhanced. The optimum yield of oil palm in peat land needs to be sustained in the medium and long term due to progressive soil subsidence. With little adoption of the integration of crops and livestock, we propose that peat land oil palm farmers in this area intensify crop or livestock integration, which is guided and supported by agriculture agencies, to diversify their income. Intensified digital dissemination of environmentally sustainable farming techniques may increase among farmers the perceived effectiveness of GAP and enhance their inclination to adopt such practices. Continuous learning may also facilitate a higher level of understanding and adoption of socio-ecologically friendly practices in peat land oil palm cultivation. In the future, organic farming technology, which is more environmentally friendly, should be gradually introduced to peat land oil palm farmers. One of the major limitations of this study is the small sample size and that the limited selected peat land area does not reflect the national level. Therefore, it may be necessary to replicate the variables in this study at a national or regional level in other peat land areas. Longitudinal assessment and review are important to ensure that sustainable agriculture programs are responsive to farmers' needs and goals.