ELEMENTS OF RURAL OFF-FARM ECONOMY: EFFECTS OF OFF-FARM INCOME ON PRODUCTIVITY AND POVERTY STATUS OF RURAL MAIZE FARMING HOUSEHOLDS IN KADUNA STATE, NIGERIA

This study evaluated elements of rural off-farm economy: effects of off-farm income on productivity and poverty status of rural maize farming households in Kaduna State, Nigeria. This research study was designed specifically to achieve the following objectives: determine the socio-economic

production frontier model.Age distributions of rural maize farming households  = 39.50 years shows that 83% of rural maize farming households were less than 50 years.About 69% of rural maize farming households had less than 10 people per household  = 6 people per household .The food poverty line was 4,210.16Naira (8.77USD).Poverty incidence ( 0 ), poverty depth ( 1 ) and poverty severity ( 2 ) of rural maize farming households were 0.517, 0.216 and 0.221 respectively.Determinants of productivity of rural maize farming households are farm size (P < 0.05), seed input (P < 0.05), fertilizer input (P < 0.05), chemical input (P < 0.10), and labour input (P < 0.05).In the inefficiency model, level of education (P < 0.05), marital status (P, 0.10), off-farm income (P < 0.05), access to extension agents (P < 0.05) and access to credit (P, 0.10) statistically and significantly reduces the probability of rural maize farming households of being food poor.Factors that statistically and significantly reduces food poverty status among rural maize farming households were gender (P < 0.05), educational level (P < 0.10), extension contact (P < 0.01) and off-farm income (P < 0.01).The retained production constraints facing rural maize farming households in the principal component analysis are lack of credit facilities, lack of fertilizers, lack of improved seeds, bad road infrastructures and lack of extension services.

KEY WORDS
Off-farm income, productivity, poverty status, rural maize farmers, Kaduna State, Nigeria.
Rural livelihoods income comes from farm, non-farm and off-farm sources.Rural farming households now diversified into off-farm and/or non-farm activities as a surviving strategy (Astatike and Gazuma, 2019).In sub-Saharan countries, off-farm income constitutes about 35-50% of income for rural farming households (Haggblade, Hazell and Reardon, 2010).Off-farm activities can be defined as activities performed by rural farmers away from his or her own farms, this involves activities performed in other peoples farms plus activities performed outside agricultural sector (Non-Farm).Non-farm activities can be defined as activities performed by rural farmers outside agricultural sector (Barrett, Reardon and Webb, 2001;Haggblade, Hazell and Reardon, 2007).
Rural off-farm economy is of great importance for rural economy because it generates employment, provides both production linkages and substantial rural income for rural households.Off-farm income relaxes the budget constraints of rural farming households.Offfarm activities improve the standard of living, livelihood security of rural farmers and absorb surplus agricultural labour input.In sub-Saharan countries, rural poverty is the major issue, majority of their population live in rural areas and basically depend on agriculture.Off-farm income would significantly add to income of rural farming households and reduce rural poverty (Iqbal et al, 2017).Poverty according to Folorunso, Gama and Ademiluyi (2018) can be defined as rural farmers' inability to adequately meet the basic human necessities such as food, clothing, shelter and medicare.According to Obayelu and Orosile (2015), food poverty is the inability of rural farming households to afford, or have access to quality food in sufficient quantity to make up a healthy diet.Food poverty among rural farmers is the result of failure of rural livelihood to guarantee access to sufficient food at the household level.Livelihood means making a living, the various activities and resources that allow people to live.Rural farming households that are food poor are also food insecure.Food insecurity and poverty are more prominent in rural areas of Nigeria.Despite the fact that rural farmers are involved in farming activities, their income remains low.
Agriculture has the potentials of reducing rural poverty and promoting economic development in Africa (Rufai, Salman and Salawu, 2018).David (2010) explained the need to improve productivity in sub-Saharan countries for it to reduce poverty and have higher impact on aggregate economic indicators.The low productivity of agriculture in sub-Saharan Africa has been attributed to the fact that the sector are made up of small-scale farmers who use rudimentary production technique and the increasing out-migration of men from rural areas and their participation in off-farm activities and left agriculture more in the hands of rural women (Lastarria-Corhiel, 2006).Low productivity among rural farmers can be further attributed to poor access to farm inputs, new technologies, credit facilities and extension services.Knowledge of efficiency of resource use among rural farmers is also vital in agricultural productivity (Edet and Falake, 2006).Productivity growth and efficiency of farm resources are core issues of sustainable agricultural production of small-scale rural farming activities.
The broad objective evaluated elements of rural off-farm economy: effects of off-farm income on productivity and poverty status of rural maize farming households in Kaduna State, Nigeria.This study was designed specifically to achieve the following objectives: (i) Identify the socio-economic characteristics of rural maize farmers; (ii) Determine the poverty status of rural maize farmers; (iii) Determine the elasticity of production and return to scale of rural maize production; (iv) Examine the effect of off-farm income on productivity of rural maize farmers; (v) Examine the effect of off-farm income on poverty status of rural maize farmers, and (vi) Determine the constraints facing rural maize farming households in the study area.

METHODS OF RESEARCH
This research study will be conducted in Kaduna State.The State lies between Latitudes 10 0 20 0 N and Longitudes 7 0 45 ‫|‬ E. Kaduna State has land mass of 46, 053Km 2 with population of 6,113,503 people (NPC, 2006).The inhabitants of the state are farmers; crops grown include maize, millet, groundnuts, cowpea, rice, yam, cassava, ginger, tomatoes, and sorghum.They also reared animals like goats, sheep, poultry birds, and cattle.Primary sources of data were used.Data were collected with the aid of well-structured questionnaire.Data were subjected to validity and reliability test.Multi-stage sampling technique was used to selected target maize farming households.First, second and third stages involves the simple random selection of Kaduna North, 5 wards, 5 villages respectively using ballot-box, raffle draw method.The fourth stage involves selecting of 100 maize farming households following Yamane (1967) formula for estimating sample size as stated thus: Where:  = Sample Size (Units), N = Sample Frame (Units), e = Level of Precision (5%).Data obtained from the field were subjected to statistical and econometrics analysis.Descriptive Statistics includes frequency distributions, mean, and percentages.This was used to have a summary statistics of data collected from maize farmers.This was used to identify the socio-economic characteristics as stated in specific objective one.
The FGT model is specified thus: Where: P αi = FGT Poverty Index for the i th Sub-Groups, N = Total Number of Maize Farmers in the Population, Y i = Per Capital Expenditure of i th Maize Farmers, Z = The Poverty Line, q = The Number of the Sampled Maize Farmers in the Population below the Poverty Line, α = The Degree of Aversion and take on the Value of 0,1,2.
The headcount ratio measures the incidence of poverty and it is obtained as: Poverty Gap: When α is equal to 1, equation becomes.
Square Poverty Gap: When α is equal to 2, the equation becomes.
About 2/3 of mean per capital household expenditure MPCHE was estimated as the poverty line, the extreme poor (those spending <1/3 of MPCHE, moderately poor (those spending <2/3 of MPCHE and the non-poor (those spending >2/3 of MPCHE).
This was used to determine the poverty status of maize farmers as stated in specific objective two (ii).
The stochastic production frontier model is stated thus: Where:  This was to examine the effect of off-farm income on productivity of maize farmers as stated in specific objective four (iv).
Return to scale of the farm operations can either be increasing, decreasing, or constant return to scale base on the value.

𝑅𝑇𝑆 =
(9) Where: RTS = Returns to Scale,   = Elasticity of Production Inputs (Units).This was used to determine elasticity of production and return to scale as stated in specific objective three (iii).
The Probit Model is stated thus: This was used to examine effect of off-farm income on poverty status of maize farmers as stated in achieve specific objective five (v).
The constraints facing rural maize farmers as stated in specific objective six (vi) were subjected to principal component analysis.Principal Component Analysis transformed many interrelated variables into few unrelated variables.

RESULTS AND DISCUSSION
The socio-economic characteristics of rural maize farmers are presented in Table 1.About 83% of rural maize farmers are less than 50 years of age.This implies that rural maize farmers were young, energetic, and resourceful in their active age.The mean age of rural maize farmers was 39 years.Also, 67% of rural maize farmers were male, while 33% were female.In terms of marital status, 37% of sampled rural maize farming households were married, while 47% were single.In addition, 89% of sampled rural maize farming households had formal education, this include: primary (30%); secondary (40%); tertiary (19%).Also, 11% of sampled rural maize farming households had no formal education.This implies that rural maize farming households have the knowledge base and can adopt innovations and research findings easily.The household sizes were large with an average of 7 people per household.About 69% of sampled rural maize farming households had less than 10 people per household.This study is in line with findings of Astatike and Gazuma (2019) who reported mean age of 44 years and average household size of 5 people per household among farming households in Southern Ethiopia.The food poverty profiles of rural maize farming households are presented in Table 2. Based on the food poverty line, households were classified into food poor households and food non-poor households.Two-third of the mean per capital households' food expenditure of the sampled rural maize farming households was used as poverty line in the same method as used by Omonona et al (2007).The poverty line was 4, 210.16Naira (8.77USD), this shows the minimum cost of eliminating poverty, the amount that could be transferred to the poor to bring their expenditure up to poverty line.The food poverty incidence (P 0 ) was 0.517, this implies that 51.70% of the rural maize farming households were poor and 49.30% were non-poor.The poverty gap index (P 1 ) usually referred to as the depth of an average poor person from the poverty line was 0.261, this implies that 26.1% of rural maize farming households were poor.The poverty severity (P 2 ) which measures the distance of each poor person to one another was found to be 0.221.This implies that among the rural maize farming 22.1% were severely poor.The mean per capital expenditure was 9,804.12 for non-poor rural maize farming households and 1,671.43 for poor rural maize farming households.This result is similar to results of Obayelu and Orosile (2015) who reported poverty incidence (P 0 ) of 0.45, poverty depth (P 1 ) of 0.25 and poverty severity of (P 2 ) of 0.16 among 150 households in Ekiti State, Nigeria.
Estimates of Stochastic production frontier model and technical efficiency scores are presented in Table 3 and 4. The coefficient of gamma (γ) of 0.69017 is very close to one, and is statistically and significantly different from zero at (P < 0.01), also Lambda (λ) of 1.3140 is greater than one and statistically significant at (P < 0.01).This suggests the presence technical inefficiency among the rural maize farmers.The results of the estimated parameters showed that all the coefficients of the technical components conform with a priori expectations of having positive signs.The statistically and significant physical variables included in the technical efficiency components are: farm size (P< 0.10), seed input (P < 0.05), fertilizer input (P < 0.05), chemical input (P < 0.05), and labour input (P < 0.05).Seed input is statistically and significantly contributes more to yields of Maize than other agricultural inputs included in the model.A 1 % increase in improved seed inputs will leads to 29.04% improvement in maize yields.Maize producers who used improved seeds per hectare performed better.These results is in line with findings of Houngue and Nonvide (2020) who showed that improved seed inputs significantly contributes 19.1% to rice production.These results also conform to findings of Oppong et al (2014) and Shabu (2013) who reported that seed inputs contributes significantly about 25% of maize and rice production in Ghana and Nigeria respectively.Chemical input contributes the least to rural maize production among farming households.A 1% increase if chemical input will result in a 5.74% improvement in maize yield.This in line with findings of Houngue and Nonvide (2020) who reported 4.8% increase in rice yield from 1% increase in herbicide input.These results are in line with those obtained by Oppong et al (2014) and Ogundari (2008).Table 3 also presents the results of the technical inefficiency model of rural maize production.The statistical and significant variables in the inefficiency model are gender (P < 0.05), age (P < 0.05), marital status (P < 0.10), level of education (P < 0.05), household size (P < 0.05), off-farm income (P < 0.05), access to extension agent (P < 0.05) and access to credit (P < 0.10).The coefficient of age is positive and significant, this implies that as rural maize farmers get older they are less efficient than younger ones, in addition they are less receptive to research findings, new innovations and technical changes in maize production that influence efficiency.The coefficient associated with level of education is negative and significant.This means level of education contributes to the reduction of inefficiency in maize production.This result is in line with previous work of Elias et al (2017) and Yabi (2009).Access to credit has negative and significant coefficient, this implies that access to credit facilities reduces the inefficiency of maize production.Access to credit facilities enables rural farmers to purchase farm inputs in time and avoid delays in farming activities (Houngue and Nonvide, 2020; Mdemu et al, 2017).Off-farm income has negative and significant coefficient.This means that off-farm income significantly increase technical efficiency and reduces inefficiency in maize production.This result is in line with previous work of Astatike and Gazuma (2019) who reported that off-farm activities significantly influence total income of households in Southern Ethiopia.The estimated technical efficiency score varies from 21% to 99% with an average of 68.3% (Table 4).About 7% of rural maize farmers have their technical inefficiencies scores less than 50%.This means there are opportunities to increase rural maize production by adopting new innovations, research findings and better farm practices.These results are similar to previous studies carried out by Houngue and Nonvide (2020), Dhungana et al (2004), Singbo and Lansink (2010), Kadiri et al (2014) and Ouedraogo (2015).
The coefficients of the physical variables included in the technical efficiency component of the stochastic production frontier model are the elasticities of production.This implies that the elasticities of production of farm size, seed input, fertilizer input, chemical input and labour input are 0.1987, 0.2904, 0.2231, 0.0574, and 0.2801respectively.The summation of elasticities of production gives the return to scale.The return to scale of rural maize production is 1.0497.This implies an increasing return to scale because the value of return to scale is greater than one.This is similar to previous result of Tesema, Kebede and Shumeta (2019) who obtained in their results the return to scale of 1.0341 in smallholder maize production in Ethiopia.The determinants of rural maize farming households food poverty status using the dichotomous Probit regression model was presented in Table 5.The Log-Likelihood value was -101.211(P < 0.01) implies that the model has a good fit.The statistically and significant explanatory variables included in the Probit model are age (P < 0.01), gender (P < 0.05), educational level (P < 0.10), household size (P < 0.05), extension contact ( P < 0.01) and offfarm income (P < 0.01).The coefficients of age and household size are positive and statistically significant in influencing food poverty status of rural maize farming households.This means as rural farming households head becomes older in age and the number of people per households increases, then the probability of being food poor increases by 1,27% and 26.37% respectively.This result is in line with previous studies of Obayelu and Orosile (2015) who reported that household food poverty increased with increase in household size and dependency ratio, which is the proportion of non-working members.The negative and statistically significant coefficient (-0.112) of educational level of the rural maize farming households head indicates that the probability of being food poor reduced by 20.31% with the household head having the highest educational attainment of tertiary education.Off-farm income has statistically and significant effect on food poverty status of rural maize farming households.Off-farm income has negative and statistically significant coefficient of -0.230, this implies that as off-farm income increases by 1% the probability of rural maize farming households being food poor reduced by 32.14%.According to Astatike and Gazuma (2019), the total annual income of households who participated in off-farm activities is found to be higher than that of non-participants, this is because households who participated in off-farm activities diversify their income sources which obviously increase their total income.Farming households with off-farm activities have higher overall income than farming households with only single source of income.According to Babatunde and Qaim (2010) households with off-farm income seems to have better access to more nutritious food, which is also reflected in significantly higher level of micronutrients consumption.Also, off-farm income increase the probability of households being food secure.Source: Field Survey (2021).*** -Significant at P < 0.01, ** -Significant at P < 0.05, * -Significant at P < 0.10.The production constraints facing rural maize farming households were subjected to principal component analysis (Table 6).Principal component analysis is an analytical tool that can translate many interrelated production variable into few unrelated ones.The retained production constraints had Eigen-value greater than one and they are lack of credit facilities (2.1672), lack of fertilizers (2.0341), lack of improved seeds (2.0021), bad road infrastructures (1.9821) and lack extension agents (1.8762).The retained components explained 84.72% of all production constraints included in the model.The Chi-square value of 520.270 was significant at (P< 0.01).
) when (α = 0) Pα = P = Poverty Incidence or Head Count Ratio Where: H = Number of Individuals below Poverty Line, N = The Number of Individuals in Reference Population.

Table 2 -
Poverty Status of Rural Maize Farming Households

Table 3 -
Stochastic Production Frontier Model for the Rural Maize Production

Table 4 -
Summary Statistics of Technical Efficiency

Table 5 -
Maximum Likelihood Estimates of Probit Regression Model

Table 6 -
Results of the Principal Component Analysis of Constraints Facing Rural Maize Farmers *** -Significant at 1% Probability Level.