Data availability
Data will be made available on request.
ECONOMIC PROFITABILITY AND EFFICIENCY OF HONEY PRODUCTION FROM THE ASIAN HONEYBEE (Apis cerana) IN THE MID-HILLS OF NEPAL
ECONOMIC PROFITABILITY AND EFFICIENCY OF HONEY PRODUCTION FROM THE ASIAN HONEYBEE (Apis cerana) IN THE MID-HILLS OF NEPAL
Keywords
modern hive
resource use
profitability
efficiency
beekeeping
honey
1. INTRODUCTION
Beekeeping is a profitable business that yields a rapid return on a relatively minimal investment, often becoming self-sustaining within its first year of operation [1]. This practice requires minimal resources and can be pursued domestically and commercially, offering supplementary income and employment opportunities, particularly for economically disadvantaged populations [2]. As a vital sub-sector, beekeeping has substantial economic potential, producing high-quality products such as honey, wax, pollen, propolis, royal jelly, and bee venom [3]. Furthermore, beekeeping contributes to agricultural productivity and environmental conservation through the natural pollination activities of bees, which enhance crop yields and support the maintenance of natural flora [4].
In Nepal, commercial beekeeping predominantly focuses on the production of honey, which is closely linked to the genetic diversity of the Apis mellifera and Apis cerana. The latter species, Apis cerana, is especially favored in Nepal’s mid-hill regions due to its use of cost-effective log hives and its greater resilience to cold climates and pests compared to Apis mellifera [5]. Beekeeping is a prominent industry in the Myagdi district, which produces 36 metric tons of honey annually, as reported by the District Agricultural Development Office, Myagdi [6]. For Apis cerana bees, local domestic farming and wild vegetation serve as primary food sources. The Myagdi district possesses significant potential for commercial beekeeping, which could substantially improve the economic conditions of participating farmers. The area’s abundant bee flora, and favorable climate conditions are conducive to commercial beekeeping [7].
The socio-economic status of beekeepers directly affects the profitability of beekeeping [2]. Currently, the average honey yield is only 3.5 kg per colony, considerably lower than the potential yield of 10 kg per colony [7]. Although farmers utilize various resources, they often do so inefficiently, raising questions about the profitability of beekeeping. Traditional practices further hinder the expansion of the beekeeping industry [8]. Access to regional productivity and profitability information is essential for adopting new technologies and modern practices in beekeeping [9]. Therefore, studying the economic profitability and efficiency of beekeeping in the Myagdi district is crucial for the development and success of this sub-sector.
Previous studies on beekeeping have not specifically focused on honey production or the beekeeping practices of Apis cerana. However, similar research has been conducted on honey production from Apis mellifera in other districts of Nepal, such as the Lamjung, Dang, Bardiya, and Chitwan districts of Nepal, but with diverse interests. The resource allocation efficiency in beekeeping was analyzed within the Chitwan district, and it was reported that resources were not allocated efficiently [10]. The beekeeping enterprises in Chitwan were profitable, reporting a benefit-cost ratio of 1.65 [11]. While all research only focuses on beekeeping's profitability and resource use efficiency. This study aimed to estimate the profitability and technical efficiency status of the beekeeping enterprises for the species Apis cerana and analyze the different determinants that can increase the adoption of modern practices in beekeeping.
The primary objective of any production system is to achieve the highest possible output with the given inputs. Measuring efficiency is essential for optimal production levels [12]. This research is motivated by the need to examine the profitability and efficiency of the beekeeping subsector in the Myagdi district of Nepal. The study compared the productivity of traditional versus modern beekeeping practices and identified factors influencing the adoption of modern practices. Profitability analysis uses parameters such as benefit-cost (B/C) ratio, gross margin, and net margin. Efficiency measurement employs methods including the Cobb-Douglas production function, the MVP-MFC approach, and the Stochastic Frontier approach. This research will provide valuable information for making informed management decisions regarding resource allocation and formulating agricultural policies and technical improvements.
Based on the reviewed literature and the objectives of this study, the following null hypotheses were formulated for testing:
H0, 1: Beekeeping enterprises in the Myagdi district utilizing modern hives and those using traditional hives do not have significantly different benefit-cost ratio (BCR) and other profitability measures.
H0, 2: Key production inputs such as labor, artificial feed, and pasture management costs are efficiently utilized.
H0, 3: The mean technical efficiency of beekeeping enterprises in the Myagdi district is at the optimal frontier.
H0, 4: The socioeconomic characteristics of beekeepers and institutional factors influence the adoption of modern hives and beekeeping technologies in the Myagdi district.
2. METHODS
2.1. Study site
This study was carried out in the Myagdi district, which lies in the mid-hills of Nepal. The map of the study site is shown in Figure 1. It extends from 28 ° 29’ 48’’ north to 28° 79 24’’ north latitude and 83° 09 82’’ east to 83° 87 27’’ east longitude. This district is stretched over mid-hill, and mountain ranges from 915 meters to 8157 meters above the average mean sea level. It has an area of 2297 sq. km and a population of 107,033 [13]. Two local bodies, Beni Municipality and Mangala Rural Municipality, are the research sites for this study.
Figure 1. Map of the study site
2.2. Sample size and Sampling technique
A multi-stage sampling method was employed, where in the first stage, Myagdi district was purposively selected from 77 districts of Nepal, as this district is noted for honeybee production for Apis cerena [7]. In the second stage, two municipalities, Beni and Mangala, were purposively selected from the major honey-producing communities, as these two municipalities are the pocket areas for honey production with a suitable environment for commercial honey production. Finally, to ensure representativeness at the household level within these pocket areas, a simple random sampling method was used to select 55 beekeepers from the list of 150 commercial beekeepers (those with more than two hives involved in selling honey) obtained from the Agriculture Knowledge Centre, Myagdi, in each municipality. The sample size was determined by Yamane’s formula [14].
Sample size = = 110, with a margin of error of 5 percent. In the final stage, 55 beekeepers from each municipality were randomly selected, and every beekeeping household had the chance of being selected.
A total of 110 beekeeping farmers were interviewed using well-structured questionnaires, focus group discussions, and key informant surveys. Data were collected for socioeconomic characteristics of beekeepers, input used for beekeeping, cost of production, output produced, and revenue from beekeeping in the years 2022 and 2023.
2.3. Data analysis
Descriptive statistics assessed honey production's economic profitability and efficiency in the study site, as well as the benefit-cost ratio, Cobb Douglas production function, resource use efficiency, and technical efficiency estimates.
2.3.1. Profitability analysis
The fixed cost of different tools used in beekeeping was calculated by the annual depreciation method. Total fixed cost = Bee colony purchase + Beehive + Queen gate + Smoker + Honey extractor + Bee veil + Knife + Beehive stand + Queen cage + Queen excluder. We divided the fixed cost by the number of equipment life expectancy years [15], as shown in Table 1.
Table 1. Number of years of life expectancy of the equipment
| Tools | Economic Life (Years) |
|---|---|
| Bee colony | 10 |
| Beehive | 10 |
| Honey extractor | 10 |
| Stand | 10 |
| Smoker | 5 |
| Knife | 5 |
| Queen gate | 5 |
| Queen cage | 5 |
| Queen excluder | 5 |
| Bee veil and gloves | 3 |
All variable inputs like human labor, sugar, drugs, comb foundation, and migration costs involved in beekeeping were considered and valued at current market prices to calculate the cost of production. During cost estimation, both purchased and owned farm-produced inputs were accounted for.
We calculated the gross return by multiplying the total volume of product from beekeeping by the average price of the product during the harvesting period. Gross margin calculation estimates the difference between the gross return and variable costs. Gross margin and benefit-cost ratio were calculated by using the following formula [15];
Gross Margin (NRs. /colony) = Gross return (NRs.1 /colony) – Total variable cost (NRs. /colony)
The benefit-cost analysis was carried out by using the formula:
2.3.2. Estimation of efficiency ratio
2.3.2.1. The Resource Use Efficiency
Resource use efficiency in honey production was estimated using the Cobb-Douglas production function. The measure of resource use efficiency was calculated by taking the ratio of Marginal Value Product (MVP) to Marginal Factor Cost (MFC) based on the regression coefficients of each input. The state of use of resources (i.e. underutilization, optimum utilization, and overutilization) was estimated by following the methodology [16].
Cobb-Douglas production function can be estimated as:Y = aX1b1 X2b2 X3b3 X4b4
Transforming to the natural logarithm(1)where Y is the total return from beekeeping NRs. /Colony, X1 is the total cost of artificial feed NRs. /Colony, X2 is the total cost of labor used in NRs. /Colony, X3 is the total cost of medicine in NRs. /Colony, X4 is the total cost for pasture management in NRs. /colony, a is Intercept, and ln represents the natural logarithm.
To calculate the return to scale, the Cobb-Douglas production function was used, and the RTS was calculated as RTS = ∑bi. The RTS measures the proportionality of output changes after a change in the quantity of all production inputs by the same factor. The returns to scale in technology are increasing, constant, and decreasing returns to scale [17]. The sum of bi from the Cobb-Douglas production function gives the value of return to scale, where bi is the coefficient of ith explanatory variable obtained from OLS regression of the Cobb-Douglas production function. The decision rule is that if RTS is less than 1, there are decreasing returns to scale; if RTS is equal to 1, there are constant returns to scale; and if RTS is more than 1, there are increasing returns to scale.
The efficiency ratio is computed by dividing the Marginal Value Product (MVP) by the Marginal Factor Cost (MFC), following the methodology outlined in [16]. This ratio, denoted as r, helps assess resource utilization. If the ratio is greater than one, it suggests that the resource is being underused, implying that increasing the input could improve efficiency. If the ratio equals one, it suggests that the resource is being used efficiently. Conversely, a value less than one signals overuse, meaning the resource is being applied beyond its optimal level.
The Marginal Value Product (MVP) is computed by multiplying the estimated regression coefficient (bi) by the ratio of the geometric mean of total revenue to the geometric mean of the respective input (Xi).
The relative percentage of change in MVP of each resource is estimated as follows:
Alternatively, using the efficiency ratio, this can be rewritten as:Here, r is the efficiency ratio, i.e., MVP divided by MFC, and D represents the absolute percentage change in the MVP of a resource, providing insight into necessary adjustments for optimal resource allocation.
2.3.2.2. The Technical efficiency
The Stochastic Production Frontier model was utilized to measure technical efficiency. Stochastic Frontier Analysis (SFA) is a parametric method that employs standard production functions, such as the Cobb-Douglas and Trans log production functions. This approach explicitly considers the maximum possible output level from a given set of inputs. The model features a production function with a two-component error term [18], [19]. One component accounts for random effects, such as measurement errors and unpredictable factors like weather or strikes, while the other component addresses technical inefficiency affecting the output variable [18]. SFA has been widely applied in numerous empirical studies and has undergone various modifications and extensions.
The SFA production function can be expressed as;(2)where = 1, 2, 3, 4, 5…. N, is the honey production per colony of the th farm, and is the production function, Xi is the inputs for the th farm, and is the coefficient estimates. ei is an error term. , νi = random effects as a result of measurement errors and other factors, which are not under control of beekeepers. This parameter is assumed to be identically, independently and normally distributed with mean zero and constant variance, N (0, )
ui = inefficiency error term, which is a non-negative random variable associated with technical inefficiency of production, assumed to be independently distributed such that ui is obtained by truncation (at zero) of the normal distribution with mean ui and variance .
From this production function, the technical efficiency of an th farm can be estimated as the ratio of the observed output of the th farm relative to the potential output defined by the frontier function [20].
Formally, the technical efficiency of th beekeeper is;(3)Here,
Major inputs in honey production are assumed to be labor, artificial feed, medicine, and a number of colonies. The following trans-log model is specified for the estimation of the Stochastic Frontier Analysis (SFA) production function. Honey production function can be represented as:
Y = aX1b1 X2b2 X3b3 X4b4 Taking log transformation,(4)where Y is the total production of honey (kg/colony), X1 is labor (man day/colony), X2 is artificial feed (kg/colony), X3 is the quantity of medicine (gram/colony), and X4 is the number of total colonies.
The technical efficiency values range from zero to one, where one represents fully efficient beekeepers. The maximum likelihood estimation (MLE) procedure was used to estimate the stochastic frontier production function.
2.3.3. Factors affecting the adoption of modern hives and technology for Beekeeping
Both modern and traditional hives for beekeeping were documented. The adopters and non-adopters of modern recommended practices were classified. Logit regression analysis was employed to find the determinants of technology adoption in beekeeping.
In this study, the dependent variable, adoption of modern hives and technology in beekeeping, is discrete, dichotomous, and mutually exclusive, followed by the binary logit model. This model contains one dependent variable with two categorical outcomes adopting modern technology in beekeeping. The prediction of whether to adopt modern technology in beekeeping is examined using ten explanatory variables. The logit model, which is based on cumulative logistic probability functions, is computationally easier to use than other models, and it also has the advantage of predicting the probability of adopting any technology [21].
This study employs a logit model to analyze the factors influencing farmers’ adoption of modern hives and beekeeping technologies. The probability of adoption, denoted as Π(X), is modeled as follows:(7)Here, Π(X) represents the probability that a farmer adopts the modern hive and technologies for beekeeping, with the dependent variable coded as 1 for adopters and 0 otherwise. The model transforms this probability into a linear relationship using the logit function:(8)
This linear form allows for easier interpretation of the relationship between explanatory variables and the log odds of adoption. The probability of adoption of modern hives and technology is influenced by a set of socioeconomic and institutional factors, which are captured in the following equation:(9)In this specification, is the probability that the th farm adopts modern beekeeping technology, 1- is the probability of non-adoption, is the logit coefficients which quantify the impacts of each explanatory variable on the log-odds of adoption, and is random disturbances for each observation (i = 1,2,3, 4,…,100). The description of the variables and their expected sign are detailed in Table 2.
Table 2. Variables used in the study’s empirical models
| Independent variable | Definition and measurement | Hypothesized sign |
|---|---|---|
| AGE | Age of the household head (years) | + |
| GEN_HHH | Gender of the household head (Male=1, Female=0) | +/- |
| EDU_STATUS | Education status (years of schooling) | + |
| ECO_ACT | No. of economically active members in the family | + |
| LIVE_HOLD | In Livestock Unit (LSU 1) | +/- |
| FOR_EMP | Outmigration (Yes=1, No=0) | +/- |
| GRP_MEM | Membership in farmer’s group (Yes=1, No=0) | + |
| TRNG | Training received (Yes=1, otherwise=0) | + |
| CRED_ACCESS | Access to credit (Easy=1, otherwise=0) | + |
| SUB_REC | Subsidies received (Yes=1, otherwise=0) | + |
- 1
- LSU: Livestock conversion unit; conversion factors: cattle (0.50), buffalo (0.50), sheep and goats (0.10), pigs (0.20), rabbits (0.02) and poultry (0.01)
(Source: FAO & AGAL, 2005, [22])
2.3.4. Diagnostic tests and model validation
Both heteroscedasticity and multicollinearity in the data are checked by calculating the Variance Inflation Factor (VIF) and Breusch-Pagan Test, respectively. For the Cobb-Douglas production function model, the VIF was 1.26, indicating no multicollinearity problem in the data. We fail to reject the null hypothesis of constant variance in the Breusch-Pagan test, indicating no heteroskedasticity problem in the data.
The Stochastic Frontier Analysis (SFA) and logistic regression model were tested for the model specification using linktest. The result shows that all the models are correctly specified and the functional form is free from misspecification.
3. RESULTS AND DISCUSSION
3.1. Socioeconomic characteristics of beekeeping households
Table 3 presents the socioeconomic characteristics of households engaged in beekeeping, contrasting traditional and modern beekeeping practices. Traditional beekeeping differs from modern methods, primarily in hive management, as bees are domesticated in tree logs and local materials are used to make hives in traditional methods. In contrast, modern hives are used in modern methods [23]. The average age of household heads was 48.30 years for modern and 49.00 years for traditional categories, ranging from 30 to 76 years. There is no significant age difference between the two groups. Education levels of household heads averaged 8.28 years, with significant variation: modern beekeepers averaged 9.61 years of education, while traditional beekeepers averaged 6.43 years, indicating a statistically significant difference at the 1% level. This suggests that higher education levels motivate the adoption of modern practices. The range of education among beekeepers was between 2 and 12 years.
Table 3. Socioeconomic characteristics of beekeeping households: summary statistics
| Variables | Overall (n = 110) | Type of hive | t-value | p-value | |
|---|---|---|---|---|---|
| Modern (n = 64) | Traditional (n = 46) | ||||
| Age (years) | 48.59 | 48.30 | 49.00 | -0.34 | 0.714 |
| Education (years) | 8.28 | 9.61 | 6.43 | 4.27*** | 0.000 |
| Household size | 5.64 | 5.72 | 5.52 | 0.58 | 0.56 |
| Male members | 2.82 | 2.95 | 2.63 | 1.65* | 0.091 |
| Female members | 2.82 | 2.76 | 2.89 | -0.58 | 0.56 |
| Economically active members | 3.95 | 4.16 | 3.67 | 1.69* | 0.093 |
| Years of beekeeping (years) | 6.32 | 6.06 | 6.67 | -0.74 | 0.45 |
| Total number of beehives with colony | 5.90 | 6.53 | 5.04 | 1.22 | 0.22 |
| Members in beekeeping | 1.12 | 1.14 | 1.10 | 0.40 | 0.68 |
| Contribution to annual income (%) | 15.77 | 16.32 | 15 | 0.64 | 0.52 |
| Livestock holding (LSU) | 1.34 | 1.18 | 1.57 | -2.10** | 0.038 |
Note: ***, ** and * indicate 1%, 5% and 10% levels of significance, respectively
Household sizes averaged 5.72 members for the modern and 5.52 for the traditional category. The average number of male members is 2.95 for the modern and 2.63 for the traditional category, while the number of female members was 2.76 for the modern and 2.89 for the traditional. Economically active family members averaged 3.95 overall, with 4.16 in the modern and 3.67 in the traditional method of beekeeping group, showing statistical significance at the 10% level. Beekeeping experience averaged 6.32 years, with 6.06 years for the modern and 6.67 years for the traditional category, showing no significant difference. The average number of beehives with colonies was 5.90, 6.53 for modern, and 5.04 for traditional.
On average, 1.12 household members were involved in beekeeping, with 1.14 in the modern and 1.10 in the traditional. Beekeeping contributed an average of 15.77% to annual household income, with approximately 16.3% for modern and 15% for traditional hive user. However, the difference in income contribution between the two groups was not statistically significant. This may be attributed to the fact that modern hive technology is predominantly adopted by households with relatively higher education levels and diversified income sources beyond beekeeping. As a result, the marginal impact of beekeeping on total income may appear diluted among modern hive adopters compared to those relying more heavily on traditional practices. These observations, with findings from similar contexts, show that beekeeping accounted for nearly 15-20% of household income in Tigray, Ethiopia [24]. Livestock holdings averaged 1.34 LSU, with 1.18 LSU in modern beekeeper households and 1.57 LSU in traditional beekeeper households.
3.2. Profitability analysis
3.3.1. Cost of Production
The cost of production per colony per year is in NRs., presented in Table 4. The total cost of honey production was NRs. 3,278.78 per colony. The total cost of honey production from A. cerana in the modern and traditional hive categories was NRs. 3,919.70 per colony and NRs. 2,387.09 per colony, respectively. The total fixed cost includes the depreciation of the bee colony, hive, instruments, and hive tools such as a stand, honey extractor, bee veils, gloves, queen gate, queen excluder, etc. Similarly, the variable cost comprised cost for various inputs such as labor, feed, medicine, marketing, repair, maintenance, etc.
Table 4. Cost of production for beekeeping per colony: summary statistics
| Particulars/Cost (NRs. /colony) | Overall (n = 110) | Type of hive | t-value | p-value | |
|---|---|---|---|---|---|
| Modern (n = 64) | Traditional (n = 46) | ||||
| Total fixed cost | 1,310.84 (39.98) | 1,475.76 (37.65) | 1,081.39 (45.30) | 9.632*** | 0.000 |
| Labor | 1,267.82 (38.67) | 1,519.72 (38.77) | 917.34 (38.43) | 10.981*** | 0.000 |
| Artificial feed | 350.05 (10.68) | 497.27 (12.69) | 145.23 (6.08) | 11.904*** | 0.000 |
| Medicine | 55.54 (1.69) | 74.44 (1.90) | 29.24 (1.22) | 2.996*** | 0.006 |
| Pasture management | 176.08 (5.37) | 188.60 (4.81) | 158.66 (6.65) | 1.081 | 0.28 |
| Repair and maintenance | 30.76 (0.94) | 36.47 (0.93) | 22.83 (0.96) | 2.425** | 0.017 |
| Marketing | 46.55 (1.42) | 56.72 (1.45) | 32.40 (1.36) | 1.575 | 0.134 |
| Comb foundation | 41.14 (1.25) | 70.72 (1.80) | 0.00 (0.00) | 4.041*** | 0.000 |
| Total variable cost | 1967.95 (60.02) | 2443.94 (62.35) | 1305.70 (54.70) | 12.238*** | 0.000 |
| Total cost | 3,278.78 | 3,919.70 | 2,387.09 | ||
Notes: *** and ** indicate 1% and 5% levels of significance, respectively; Figures in parentheses indicate a percentage of the category.
Among the variable costs, the cost of labor, the cost of artificial feed, the cost of medicine, and the cost of comb foundation were highly significant at the 1% level among modern and traditional hive categories. The repair and maintenance cost was significant at 5%, and the total variable cost was highly significant at 1%. Meanwhile, marketing and pasture management costs were found to be insignificant, i.e., modern and traditional hive categories have uniform costs on respective items, as shown in Table 4.
3.2.2. Production and productivity of honey and wax
In 110 sampled beekeepers, on average, 20.16 kg of honey was produced by beekeepers in the modern hive and 10.72 kg in the traditional hive, as shown in Table 5. Similarly, productivity was 3.15 and 2.10 kg per colony of the modern and traditional hive categories. The mean difference in the production and productivity of honey of Apis cerana is significant at the 1% level.
Table 5. Production and productivity of honey and wax per colony: summary statistics
| Particulars | Overall (n = 110) | Type of hive | t-value | p-value | |
|---|---|---|---|---|---|
| Modern (n = 64) | Traditional (n = 46) | ||||
| Production (kg) | 16.21 | 20.16 | 10.72 | 3.30*** | 0.004 |
| Productivity (kg/colony) | 2.71 | 3.15 | 2.10 | 9.41*** | 0.000 |
| Wax (g /colony) | 5.80 | 9.97 | 0.00 | 2.06* | 0.08 |
Note: *** and * indicate 1% and 10% levels of significance, respectively.
3.2.3. Revenue from honey and other sources
In the modern hive category, NRs. 1,618.19 per colony was received from selling colonies and other products in one year, whereas only NRs. 473.79 was received in the traditional hive category, shown in Figure 2. Most returns from beekeeping were obtained from selling honey compared to other products.
Figure 2. Revenue from beekeeping by type of hive
3.2.4. Benefit-cost ratio
The average total cost/hive was NRs. 3,278.79, which gives the return of NRs. 5,477.07. Profitability analysis of beekeeping showed that the average BCR was found to be 1.67. The BCR for the modern and traditional hive categories was 1.73 and 1.61, respectively, as shown in Table 6. Similarly, gross and net margins were NRs. 3,509.12 and NRs. 2,198.28, respectively. The gross margin was found to be NRs. 4,214.25 and NRs. 2,528.09 in the modern and traditional hive categories, respectively, as shown in Table 6. The mean difference in gross margin and net margin is statistically significant at a 1% level for modern and traditional hive categories. However, the mean difference in BCR is statistically significant at 10% for both modern and traditional hive categories. The profitability analysis of beekeeping in the Myagdi district reveals it to be a low-cost, high-return enterprise for farmers. These findings align with similar studies conducted in other regions, which consistently report a BCR greater than one [10], [11], [25]. Modern beehives significantly improve honey production and facilitate easier colony division compared to traditional hives, with nearly double returns. This finding aligns with the study in Saudi Arabia [9], which also reported a higher BCR for modern hives compared to traditional ones.
Table 6. Benefit-Cost ratio of beekeeping: summary statistics
| Particulars | Overall (n = 110) | Type of hive | t-value | p-value | |
|---|---|---|---|---|---|
| Modern (n = 64) | Traditional (n = 46) | ||||
| Total fixed cost | 1,310.84 | 1,475.76 | 1,081.39 | 9.63*** | 0.000 |
| Total variable cost | 1,967.95 | 2,443.94 | 1,305.70 | 12.24*** | 0.000 |
| Total cost | 3,278.79 | 3,919.70 | 2,387.09 | 13.23*** | 0.000 |
| Total benefit | 5,477.07 | 6,658.19 | 3,833.79 | 8.32*** | 0.000 |
| Gross margin | 3,509.12 | 4,214.25 | 2,528.09 | 5.62*** | 0.000 |
| Net margin | 2,198.28 | 2,738.48 | 1,446.70 | 4.26*** | 0.000 |
| BC ratio | 1.67 | 1.73 | 1.61 | 1.81* | 0.07 |
Note: *** and * indicate 1% and 10% levels of significance, respectively
To further explore the factors associated with greater profitability among beekeepers, we estimated a binary logistic regression model. The dependent variable equals 1 if the producer’s BCR was greater than the 1.67 average and 0 otherwise. The independent variables include demographic, management, and institutional variables. Table 7 reports the estimated coefficients, standard errors, and marginal effects. Among the included variables, livestock holding and receipt of subsidies indicate higher levels of profitability. When examining the marginal effects, the probability of achieving above-average profitability is reduced by 18 percent with additional units of livestock, while the probability of above-average profitability increases by 28 percent with a subsidy received.
Table 7. Results of binary logistic regression for determinants of high BCR among beekeepers
| Variables | Coefficients | Std. error | Z | p-value | dy/dx |
|---|---|---|---|---|---|
| Gender # | 0.28 | 0.59 | 0.48 | 0.63 | 0.05 |
| Years of education | 0.02 | 0.07 | 0.26 | 0.80 | 0.01 |
| Years of beekeeping | 0.05 | 0.05 | 0.91 | 0.36 | 0.01 |
| Livestock holding | -0.96 | 0.35 | -2.78 | 0.00 | -0.18*** |
| Training # | 0.44 | 0.61 | 0.72 | 0.47 | 0.08 |
| Subsidies received # | 1.41 | 0.57 | 2.50 | 0.01 | 0.28** |
| Types of beehive # | -0.39 | 0.59 | -0.65 | 0.51 | -0.07 |
| Constant | -0.79 | 0.81 | -0.97 | 0.33 | |
| Notes: *** and ** indicate 1% and 5% levels of significance, respectively. # represents a dummy variable. | |||||
| No. of observations | 110 | ||||
| Log-likelihood | -59.59 | ||||
| LR Chi-squared | 29.66 | ||||
| Prob>chi2 | 0.00 | ||||
| Pseudo R- squared | 0.19 | ||||
The significant negative coefficient for livestock holding suggests that higher livestock ownership reduces the probability of a beekeeper attaining above-average profitability. Households with substantial livestock may diversify their attention and labor resources away from beekeeping, reducing profitability [27]. Access to subsidies and financial support has been shown to improve the smallholder beekeepers’ profitability and technology adoption [28].
3.3. Efficiency analysis
3.3.1. Resource productivity
The Cobb-Douglas production function was employed to examine the effect of various inputs on total revenue per colony. Estimated elasticity values and related statistics of the Cobb-Douglas production function employed are presented in Table 8. The estimated coefficients represent the elasticity of each input. All four factors significantly influenced revenue per colony, as depicted by the p-value. Labor costs, artificial feed costs, and pasture management costs have a positive influence, while medicine costs negatively influence revenue per colony of beekeeping. All of these elasticities, except for medicine cost, are found to be significant. The sum of all regression coefficients yields the return to scale value of 0.84. Productivity analysis shows that a 100 percent increase in labor cost increased revenue per colony by nearly 60 percent. In comparison, a similar increase in artificial feed cost yielded only a 2 percent increase in revenue. Likewise, a 100 percent increase in pasture management costs led to a 21 percent increase in revenue per colony. According to the study [29], best management practices require an increase in labor and material costs, which results in an increase in the colony's productivity and survival. The return to scale value for beekeeping in the study area was 0.84. This implies decreasing return-to-scale conditions; an increase in all inputs by 100 percent will increase the revenue by 84 percent.
Table 8. Results of Cobb-Douglas productivity of resources used in beekeeping
| Factors/ Costs (in log) | Coefficient | Std. Error | t-value | p-value |
|---|---|---|---|---|
| Labor | 0.607*** | 0.088 | 6.89 | 0.000 |
| Artificial feed | 0.028* | 0.016 | 1.74 | 0.084 |
| Pasture management | 0.210*** | 0.052 | 4.01 | 0.000 |
| Medicine | -0.003 | 0.012 | -0.26 | 0.79 |
| Constant | 3.042*** | 0.573 | 5.35 | 0.000 |
| F-value | 32.83 | 0.000 | ||
| R2 | 0.55 | |||
| Adjusted R2 | 0.54 | |||
| Returns to scale | 0.84 |
Note: *** and * indicate 1% and 10% levels of significance, respectively
3.3.2. Resource Use Efficiency
Each input’s Marginal Value Product (MVP) is calculated based on the geometric mean (GM), the respective input coefficient, and GM or revenue per colony. The efficiency ratio (r) is calculated based on the value of MVP and MFC. The state of the resource use of each factor and the adjustment required are determined by using the efficiency ratio, as represented in Table 9. The resource use efficiency analysis shows that resources, i.e., labor, artificial feed, and pasture management, were underutilized, and there is scope for improvement by 60, 27, and 86 percent, respectively. Similar studies in other districts of Nepal, such as Lamjung, Dang, and Chitwan, also report suboptimal resource allocation. For instance, a study in the Chitwan district [10] examined the allocative efficiency of resource use in beekeeping. Their findings are quite similar to this study’s findings and indicate that inputs such as human labor, expenditure on sugar, drugs, comb foundation, and migration costs significantly contributed to beekeeping productivity, and these inputs needed to be increased by 39, 34, and 74, respectively, to achieve optimal returns. To enhance revenue per colony, beekeepers in Myagdi should prioritize efficient use of labor, artificial feed, and pasture management.
Table 9. Estimates of the resource use efficiency analysis of beekeeping
| Cost (NRs. /hive) | Geometric Mean | Regression Coefficient | MVP | MFC | r = MVP/MFC | Efficiency | Adjustment Required (%) |
|---|---|---|---|---|---|---|---|
| Labor | 1190.21 | 0.60 | 2.52 | 1 | 3.25 | Underutilized | 60.31 |
| Artificial feed | 101.70 | 0.028 | 1.38 | 1 | 1.38 | Underutilized | 27.30 |
| Medicine | NS | ||||||
| Pasture management | 142.82 | 0.21 | 7.26 | 1 | 7.26 | Underutilized | 86.22 |
Note: NS indicates Not Significant
3.3.3. Technical efficiency
Table 10 presents the maximum likelihood estimates for the parameters of the stochastic production frontier. All variables were mean corrected for normalization, enabling their interpretation as partial elasticities. The positive sum of the model’s coefficients confirms the satisfaction of the monotonicity condition.
Table 10. Maximum likelihood estimates of the stochastic production frontier model for technical efficiency
| Parameters | Coefficient | Std. error | z | p-value |
|---|---|---|---|---|
| Labor (man day/colony) | 0.37*** | 0.07 | 5.63 | 0.00 |
| Artificial feed (kg/colony) | 0.02* | 0.01 | 1.70 | 0.09 |
| Medicine(gram/colony) | -0.01 | 0.01 | -1.52 | 0.13 |
| Pasture management (NRs. /colony) | 0.12*** | 0.4 | 2.96 | 0.00 |
| Number of colonies | 0.08 | 0.05 | 1.40 | 0.16 |
| Constant | -0.20 | 0.22 | -0.92 | 0.36 |
| LN sigma2v | -3.76*** | 0.64 | -5.87 | 0.00 |
| LN sigma2u | -2.62*** | 0.66 | -4.00 | 0.00 |
| Sigma_v | 0.15 | 0.04 | ||
| Sigma_u | 0.27 | 0.09 | ||
| Sigma2 | 0.10 | 0.04 | ||
| Lambda | 1.77 | 0.13 | ||
| No. of observations | 110 | |||
| Wald Chi-squared | 94.36 | |||
| Prob>Chi-squared | 0.0000 | |||
| Log-likelihood | 10.18 |
Note: *** and * indicate 1% and 10% levels of significance, respectively
The study indicates that the labor (man-day per colony), artificial feed (kg per colony), and pasture management cost (NRs. per colony) significantly impact the honey (kg per colony) output level in the study area. From the table, the coefficients of labor and pasture management are 0.417 and 0.11, respectively, which are statistically significant at the 1% level. Likewise, the coefficient for artificial feed is 0.02, which is statistically significant at a 10% level. However, the number of colonies and medicines (grams per colony) are not significant.
Table 11 presents the frequency distribution of beekeepers in the study area, revealing a range of technical efficiency from 56 to 94 percent.
Table 11. Frequency distribution of the technical efficiency index
| Range | Frequency | Percentage (%) |
|---|---|---|
| Less than 0.60 | 2 | 0.02 |
| 0.61 to 0.70 | 14 | 12.96 |
| 0.71 to 0.80 | 24 | 22.22 |
| 0.81 to 0.90 | 52 | 48.15 |
| 0.91 to 0.99 | 18 | 16.67 |
| Total | 110 | 100.00 |
| Mean = 0.82 | Standard deviation = 0.01 | |
| Minimum = 0.56 | Maximum = 0.94 |
Figure 3 represents the kernel density plot of technical efficiency, revealing that the mean efficiency level is approximately 0.81. While a significant portion of beekeepers operate above this average, a notable share remains below, indicating potential for improvement. This efficiency gap highlights the need for targeted services such as training, subsidies to make them efficient and increase the honey production.
Figure 3. Kernel density plot of the technical efficiency of the farms
Using the Stochastic Frontier Analysis (SFA) approach, the study revealed a lambda value greater than one, indicating that inefficiencies contribute more to output variation than random errors. A study reported lambda values of 1.06 in white honey production farms in Egypt [26], suggesting that the impacts of technical inefficiencies exceed the random error factor. This implies that beekeepers are not optimally utilizing available production factors. The results indicated that 16 percent of the beekeepers are near the technical efficiency frontier. However, even the most efficient beekeepers do not optimally allocate resources and require improvements to reach the technical efficiency frontier. The mean technical efficiency of beekeepers in the study area is 0.81, suggesting that, on average, beekeepers achieve 81% of the potential output. This suggests a 19% gap in efficiency that could be bridged through better management practices without additional resource inputs. Comparative studies have reported technical efficiency levels between 78 and 86 percent for white honey production in Egypt [26]. Other similar studies identified a technical efficiency of 75 percent in stingless beekeeping in Malaysia [30] and 77 percent in beekeeping enterprises in Southern Ethiopia [31].
A binary logit regression model was estimated to further examine the factors associated with higher technical efficiency in beekeeping operations. The dependent variable equals 1 if a producer’s technical efficiency score was greater than or equal to the average of 0.8155 and 0 otherwise. The independent variables included demographic characteristics, production experience, and institutional variables. Table 12 reports the estimated coefficients, standard errors, and marginal effects. Livestock holding, subsidies, and types of beehives were shown to have an impact on higher technical efficiency. With an additional unit of livestock holding, the probability of achieving high technical efficiency decreases by 15 percent if beekeepers receive subsidies or use modern bee hives, and the probability of achieving high technical efficiency increases by 19 percent, respectively. Over-diversification into livestock without specialization may influence the resource allocation, thereby reducing the efficiency in other specific enterprises, i.e., beekeeping [27]. Access to subsidies enhances input affordability and results in efficient use of resources [28]. Modern hives enhance productivity and colony management as modern equipment is linked to higher technical performance in small-scale agricultural settings [32].
Table 12. Logit regression results of determinants of high technical efficiency among beekeepers
| Variables | Coefficients | Std. error | Z | p-value | dy/dx |
|---|---|---|---|---|---|
| Gender # | -0.38 | 0.58 | -0.65 | 0.52 | -0.07 |
| Years of education | -0.02 | 0.07 | -0.34 | 0.74 | -0.01 |
| Years of beekeeping | -0.02 | 0.05 | -0.31 | 0.76 | -0.01 |
| Livestock holding | -0.77 | 0.27 | -2.91 | 0.00 | -0.15*** |
| Training # | -0.71 | 0.58 | -1.22 | 0.22 | -0.13 |
| Subsidies received # | 0.95 | 0.53 | 1.80 | 0.07 | 0.19* |
| Types of bee hive # | 0.94 | 0.54 | 1.74 | 0.08 | 0.19* |
| Constant | 1.39 | 0.82 | 1.69 | 0.09 | |
| Notes: *** and * indicate 1% and 10% levels of significance, respectively. # represents a dummy variable. | |||||
| No. of observations | 110 | ||||
| Log-likelihood | -63.07 | ||||
| LR Chi-squared | 22.70 | ||||
| Prob>chi2 | 0.00 | ||||
| Pseudo R- squared | 0.15 | ||||
3.4. Factors affecting the adoption of modern hives and technology for beekeeping
We employ the logit regression model to determine the factors influencing the adoption of modern hives and technology for beekeeping. The estimated coefficients, marginal effects, and associated statistics of the model are detailed in Table 13. The analysis presented in Table 13 indicates that the household head’s years of education, membership in farmers’ groups, and training from the government positively and significantly influence the adoption of modern hives and technology. The gender dummy has a negative impact on the adoption.
Table 13. Logit regression results of factors affecting the adoption of modern hives and technology for beekeeping
| Variables | Coefficients | Std. error | Z | p-value | dy/dx |
|---|---|---|---|---|---|
| Age | 0.03 | 0.289 | 1.27 | 0.205 | 0.008 |
| Gender # | -2.11 | 0.739 | -2.86 | 0.004 | -0.373*** |
| Years of education | 0.22 | 0.088 | 2.46 | 0.014 | 0.049** |
| Economically active members | -0.05 | 0.185 | -0.27 | 0.786 | -0.011 |
| Livestock holding | -0.21 | 0.278 | -0.77 | 0.443 | -0.048 |
| Outmigration # | 0.22 | 0.574 | 0.40 | 0.692 | 0.051 |
| Membership in farmer’s group # | 1.31 | 0.612 | 2.15 | 0.032 | 0.292** |
| Training # | 1.98 | 0.638 | 3.10 | 0.002 | 0.440*** |
| Credit access # | -1.05 | 0.660 | -1.60 | 0.110 | -0.231 |
| Subsidies received # | 0.74 | 0.635 | 1.17 | 0.242 | 0.169 |
| Constant | -2.59 | 1.984 | -1.31 | 0.192 | |
| Notes: *** and ** indicate 1% and 5% levels of significance, respectively. # represents a dummy variable. | |||||
| No. of observations | 110 | ||||
| Log-likelihood | -44.69 | ||||
| LR Chi-squared | 60.15 | ||||
| Prob>chi2 | 0.000 | ||||
| Pseudo R- squared | 0.40 | ||||
The predominance of foreign employment in Myagdi means that females are primarily involved in training and production processes. The study found that an additional year of education for the household head increases the likelihood of adopting recommended modern technology in beekeeping by 4.9%. This is attributed to the fact that schooling improves skills and motivates the adoption of modern practices. Training increases the likelihood of adoption by almost 44%. The result shows that membership in a farmer's group increases the likelihood of adoption by almost 30%. The adoption of modern beekeeping practices by the male household heads is 37.3 % less than by the female household heads. A similar study in Southern Ethiopia [33] identified that five out of fifteen explanatory variables significantly impact the adoption of modern hive technology. These variables include the education level of respondents, land size, extension contacts, access to credit, and market accessibility. Additionally, another study in Oromia [34] found that participation in nonfarm activities, farmyard land size, beekeeping training, type of house owned, participation in beekeeping demonstrations, and the square root of landholding significantly and positively influence the adoption of modern hives.
The government support for input subsidies could significantly reduce costs and increase beekeepers’ net returns, highlighting the area’s high economic potential. The primary variable cost in honey production was labor, consistent with findings in Chitwan, Nepal [15]. Low productivity in the study area was attributed to off-seasonal rainfall, erratic weather conditions, and pest and disease attacks [32], [36]. Despite these challenges, honey production from Apis cerana is profitable across all production scales. However, commercialization and adopting new technologies, coupled with policies focused on training and input optimization, could further enhance efficiency and farm returns. It is also necessary to consider the factors that are responsible for the adoption of modern hives and technology.
4. CONCLUSIONS
This study aims to estimate beekeeping's profit indicators and efficiency in the study area. This study demonstrates that beekeeping in the Myagdi district is a profitable enterprise with significant potential for improvement through modern practices. While the average BCR of 1.67 confirms its economic profitability, there exist inefficiencies in resource utilization, particularly in labor, artificial feed, and pasture management. The technical efficiency value of 0.81 suggests that honey production can be increased by 19 percent through better management practices with given inputs. Modern beekeeping techniques and better resource management are needed to enhance the profitability and sustainability of the beekeeping industry in Myagdi. Education, training, and government support are the positive determinants of adopting modern technology in beekeeping.
The findings suggest several key policy directions. Government policies and support programs focusing on input subsidies, training, and access to modern equipment could significantly improve efficiency and profitability for beekeepers. Enhancing access to a comprehensive training program on modern beekeeping techniques, resource management, and business skills is crucial. Furthermore, promoting and strengthening beekeeper groups ao cooperatives can facilitate knowledge sharing, improve bargaining power, and ease access to modern equipment and markets. Given that female-headed households show better adoption of modern practices, policies should also aim to empower women in beekeeping further, potentially through tailored support and resources.
This study, while providing valuable insights into the beekeeping sector in the Myagdi district, has certain limitations. The cross-sectional nature of the data provides results for only one time and does not capture dynamic changes over multiple seasons or years. The study relied on beekeeper recall for some data, which may be subject to error. Additionally, while key inputs were analyzed, unobserved factors such as micro-climatic variations within the district or specific pest and disease incidences at the individual farm level could influence productivity and efficiency. Future research should aim to address these constraints, perhaps through longitudinal studies. Further investigation into the specific barriers to accessing credit, the impact of climate change on local bee flora, and a deeper analysis of the value chain for honey and other bee products from the Myagdi district would also provide a more comprehensive understanding and help maximize the sector's potential.
CRediT authorship contribution statement
Anup Paudel: Writing – original draft, Visualization, Resources, Methodology, Formal analysis, Data curation, Conceptualization. Narayan Prasad Tiwari: Visualization, Validation, Supervision, Methodology, Conceptualization. Suryamani Dhungana: Supervision, Resources, Project administration, Methodology. Logan Britton: Writing – review & editing, Validation, Supervision, Methodology, Formal analysis
Uncited reference
[35].
Declaration of interest statement
There are no competing interests.
Declaration of generative AI and AI-assisted technologies in the writing process
While preparing this work, the authors used ChatGPT to improve the grammar and language. After using this tool, the authors reviewed and edited the content as needed and took full responsibility for the publication's content.
Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Declaration of Competing Interest
☒ The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgement
The authors express their deepest gratitude to all the beekeepers and respondents for the survey. The authors want to acknowledge the Department of Agricultural Economics and Agribusiness Management, Agriculture and Forestry University, Rampur, Chitwan, for administrative support.
대화 참여하기