Assessing the Impact of Privatizing Public Agricultural Extension Services on Smallholder Farmers’ Performance: A Case Study of Thulamela and Collins Chabane Municipalities, South Africa

Thulamela and Collins Chabane Municipalities serve as representative case studies of smallholder farming conditions prevalent in rural South Africa, particularly within Limpopo Province. These municipalities share common characteristics with many other rural agricultural regions in the country, including a predominance of small-scale and subsistence farming, mixed cropping systems, and reliance on rain-fed agriculture (Lidshegu and Kabanda, 2022; Mapiye et al., 2021). Socio-economic challenges such as limited access to market infrastructure, extension services, and financial resources further align these municipalities with broader trends affecting smallholder farmers across South Africa. Therefore, insights drawn from this study offer valuable implications for understanding the impact of agricultural extension service privatization in similar rural contexts nationwide (Lidshegu and Kabanda, 2022; Mapiye et al., 2021).

The study was conducted in the Vhembe District of Limpopo Province, South Africa, specifically in Thulamela and Collins Chabane Municipalities. Vhembe District, covering 21,407 square kilometers, is characterized by a subtropical climate, with temperatures ranging from 10°C in winter to 40°C in summer and an annual rainfall of approximately 500 mm, primarily occurring between October and March. The district’s soil varies from sandy in the west to loamy and clay-rich in the east, with low inherent fertility. Agriculture in the region is divided into large-scale commercial farming and smallholder farming, with smallholder farmers primarily producing vegetables, maize, and other subsistence crops on plots averaging 1.5 hectares.

Fig. 1.

Study site, Vhembe District, Limpopo Province

Source: own elaboration, 2022.

Data were collected from March to August 2022 using a structured questionnaire administered to a stratified random sample of 319 smallholder farmers. A pre-tested questionnaire was the primary data collection tool, administered through face-to-face interviews by trained enumerators. The questionnaire was translated into the local language spoken by the farmers to ensure comprehension and accuracy. Before data collection, a stratified random sampling approach was applied to categorize farmers based on their municipal affiliation within the Vhembe District. The population was divided into two sub-groups, with 176 farmers from Collins Chabane Municipality and 404 from Thulamela Municipality.

The sample size determination for the study was computed based on the formula: (1) N=NNe2 N = \frac{N}{{N{e^2}}} where:

• n – is the desired sample size;

• N – is the total target population;

• e – the degree of accuracy required, normally set at 0.05 (5% of acceptable sampling error) (Kothari, 2004; Asfaw et al., 2017).

Based on this calculation, 319 smallholder farmers were randomly selected, with 121 participants from Collins Chabane Municipality and 198 from Thulamela Municipality.

The Statistical Package for Social Scientists (SPSS) version 26 and stataSE 17 software were used to collect and analyze the data, respectively. To analyze the impact of privatized agricultural extension services on smallholder farmers’ performance, this study employed a combination of descriptive statistics and econometric modeling, specifically a Multiple Linear Regression (MLR) model.

Mean comparisons (T-tests) were used to evaluate income differences between farmers using public vs. privatized extension services. One-way ANOVA assessed whether significant differences exist in farm income across different levels of household size, education, and market access.

This study employs a Multiple Linear Regression (MLR) model to assess the impact of privatized agricultural extension services on smallholder farmers’ performance. MLR is widely used in agricultural economics to analyze how multiple factors such as extension services, credit access, market access, gender, and education affect farm income (Gujarati and Porter, 2009; Feder et al., 2010). It estimates marginal effects, providing insights into the economic benefits of privatized extension (Davis et al., 2010). The model accommodates both continuous and categorical variables, controls for external shocks, and directly tests whether privatized extension improves farm income more than public services (Aker, 2011)

The regression equation used in this study is as follows: Net annual income from farmers'production (NAIFP)=β0+β1(Age)+β2(Gender)+β3(Marital status)+β4(Education level)+β5(Household size)+β6(Landsize)+β7(Credit access)+β8(Market access)+β9(Public extension access)+β10(Private extensionaccess)+β11(Farming experience)+β12(Extensionfeedback)+β13(Climate adaptation practices)+β14(Farm input)+β15(Market price information access)+β16(Membership in farmer cooperatives)+μi \[\begin{array}{*{20}{c}} {{\rm{Net annual income from farmers}}'\,{\rm{production (NAIFP)}}}\\ { = {{\rm{\beta }}_0} + {{\rm{\beta }}_1}({\rm{Age}}) + {{\rm{\beta }}_2}({\rm{Gender}}) + {{\rm{\beta }}_3}({\rm{Marital status}}) + }\\ {{{\rm{\beta }}_4}({\rm{Education level}}) + {{\rm{\beta }}_5}({\rm{Household size}}) + {{\rm{\beta }}_6}({\rm{Land}}}\\ {{\rm{size}}) + {{\rm{\beta }}_7}({\rm{Credit access}}) + {{\rm{\beta }}_8}({\rm{Market access}}) + }\\ {{{\rm{\beta }}_9}({\rm{Public extension access}}) + {{\rm{\beta }}_{10}}({\rm{Private extension}}}\\ {{\rm{access}}) + {{\rm{\beta }}_{11}}({\rm{Farming experience}}) + {{\rm{\beta }}_{12}}({\rm{Extension}}}\\ {{\rm{feedback}}) + {{\rm{\beta }}_{13}}({\rm{Climate adaptation practices}}) + }\\ {{{\rm{\beta }}_{14}}({\rm{Farm input}}) + {{\rm{\beta }}_{15}}({\rm{Market price information access}}}\\ {) + {{\rm{\beta }}_{16}}({\rm{Membership in farmer cooperatives}}) + {{\rm{\mu }}_i}} \end{array}\] Where:

• Y_i – net annual income from farming (dependent variable, measured in South African Rand)

• β₀ – intercept

• X₁, X₂, …, X_n – independent variables representing farmer characteristics, market factors, and extension service types

• β_i = estimated parameters indicating the effect of each independent variable on farm income

• μ_i = disturbance term capturing unobserved factors

The study surveyed 319 smallholder farmers from Collins Chabane and Thulamela Municipalities. The socioeconomic profiles of the 319 smallholder farmers who participated in the study are shown in Table 2.

The sample consisted of 59.9% female and 40.1% male farmers, indicating a significant representation of women in smallholder farming. The majority of the respondents were aged 46–55 years (27.6%), followed by those aged 66 years and older (24.5%), suggesting that smallholder farming is predominantly practiced by middle-aged and elderly individuals. Regarding education, 45.8% had completed secondary school, 23.8% had primary education, 19.4% had tertiary education, while 11% had never attended school. This suggests that a substantial proportion of smallholder farmers have at least basic literacy, which may influence their ability to adopt new agricultural technologies and practices. Additionally, 56.5% of households had between 1 to 5 members, while 41.8% had 6 to 10 members, highlighting that family labor remains a crucial component of smallholder farming in the study area.

The t-test analysis assessed the relationship between key determinants and annual farm income. The results of the t-test analysis of the determinants of smallholder farmers’ performance are shown in Table 3.

The results reveal that farmers who had access to privatized extension services earned significantly higher annual incomes (R24,570.59) compared to those relying solely on public extension services (R19,417.04) (p < 0.05). Timely extension feedback significantly influenced income levels, with farmers who received prompt advisory services earning R21,195.52, while those experiencing delays earned R12,610.02 (p < 0.01). Market access played a crucial role in improving income, with farmers having direct market linkages earning R19,518.77, compared to R14,216.49 for those lacking market access (p < 0.01). Membership in farmer cooperatives was associated with significantly higher farm incomes, suggesting that collective action strengthens farmers’ bargaining power and market participation (p < 0.05).

The one-way ANOVA results highlighted additional socioeconomic factors influencing farm performance, as shown in Table 4.

The results revealed that household size was significantly associated with income levels (p < 0.01), with larger households (6–10 members) earning more income than smaller households. Transparency and accountability in extension services also played a role, as farmers who disagreed with claims of poor transparency had higher annual incomes (R26,946.09) compared to those who strongly agreed (R13,777.84) (p < 0.01). Climate adaptation practices were significant (p < 0.05), indicating that farmers who adopted climate-smart techniques registered higher farm incomes.

The results of the multiple linear regression analysis of the impact of privatized agricultural extension services on smallholder farmers’ performance are presented in Table 5.

Privatized extension services positively influenced farm income (p < 0.05), with farmers earning R7,663.78 more than those relying on public extension. Public extension access had a negative impact, reducing farm income by R3,320.07, and highlighting inefficiencies in the public sector (p < 0.05). Timely extension feedback significantly increased earnings (p < 0.01), demonstrating the importance of real-time agricultural advisory support. Farm input access and market price information access positively affected income, showing that farmers with reliable access to inputs and price intelligence had better financial outcomes (p < 0.05). Membership of farmer cooperatives increased earnings by R2,703.57, reinforcing the benefits of collective action (p < 0.10). Climate adaptation practices also had a significant impact, indicating that resilient farming strategies contribute to higher productivity and profitability (p < 0.05).