Determinants of food insecurity in rural households: A case of lemo district, southern Ethiopia

Abstract

Despite the increasing global concern of improving food security, the determinants of food insecurity at household level in the rural areas have been poorly known. Therefore, the aim of this study was to analyze determinants of food insecurity at household level. A total of 383 households were selected using multistage sampling techniques. The logistic regression model was used to analyze the data. The result revealed that odds of illiterate households were 2.376 times more likely than educated households to experience food insecurity in the rural areas (ref. (Coef. = 0.865, OR = 2.376, P = 0.006)). Households with landholdings of more than half hectare were less likely to experience food insecurity. (ref. (Coef. = 1.982, OR = 7.260, P = 0.000)). Odds of households who engaged in off-farm activities were 0.204 times less likely to experience food insecurity. (ref. (Coef. = −1.588, OR = 0.204, P = 0.000)). Households who adopt farm technologies were less likely to experience food insecurity than those who do not adopt farm technology (ref. (Coef. = −1.086, OR = 0.337, P = 0.001)). Odds of higher-aged household heads were 6.141 times more likely to experience food insecurity than younger-aged household heads (ref. (Coef. = 1.815, OR = 6.141, P = 0.000)). Larger household sizes were less likely to experience food insecurity (ref. (Coef. = −2.423, OR = 0.089, P = 0.000)). In conclusion, understanding determinants of food insecurity at household level is essential to achieve food security in rural areas. Results suggest implementation of the effective developmental programs are needed to reduce food insecurity in rural areas.

1. Introduction

Food insecurity is a global hazard that is threatening every country of the world. Nearly 124 million people in 51 countries and territories experienced severe food insecurity at crisis levels or worse in 2017 [1]. For this reason, food insecurity attracted attentions of many program implementers, policy-makers, researchers, and decision-makers [2,3].

Food security is referred to as a situation that exists when all people, at all times, have physical, social and economic access to sufficient, safe and nutritious food that meets their dietary needs and food preferences for an active, and healthy life [4,5]. In contrast, food insecurity is a situation when people are unable to access sufficient, safe, and nutritious food and food preferences for healthier life [6]. Food security is dependent on agricultural production, food imports, employment, policy, decision-making of farmers, health access, subsidy, and resource allocation [7]. Additionally, food security would have been achieved by adopting sustainable on-farm trees integration into the farms to boost agricultural productivity and biodiversity conservation [8].

Food insecurity is an urgent global challenge [9,10], particularly in the developing world remains an important economic challenge because of information scarcity on the severity, and causes of food insecurity [3,11]. The Millennium Development Goal on poverty and hunger aimed at reducing the number of undernourished people by 50% until 2015 worldwide [12]. However, the figure of undernourished people was reached 1.02 billion worldwide in 2009 [13]. There are various anthropogenic factors and natural disasters mainly causing food insecurity. For instance, political instability, war, and natural disasters are the major causes of food insecurity in Africa where about one billion people estimated to be undernourished [14]. Achieving food security remains challenges in many rural dwellers of Sub-Saharan Africa [1,15]. Ethiopia is one of the Sub-Saharan countries facing food insecurity with more than 50% of the total population whose livelihood is based on agriculture [1,16].

Agriculture is the mainstay of Ethiopian economy [17,18], which contribute to 50% of gross domestic product (GDP). The agricultural sector dominated by smallholder farming systems that is facing constraints including shrinking land sizes, limited resources, reducing soil quality, and climate change that affect sustainable food security and agricultural production [18]. In addition, vegetation losses are exacerbated by population pressure and deforestation resulting the crisis of production, market, policy, institutional and organizational failures [19]. Thus, in order to improve a sustainable food security, a better understanding of food insecurity drivers and impacts at the rural household level is needed.

[20] pointed out considerable investments are made to alleviate poverty and ensure food security in Sub-Saharan Africa. Land degradation, land shortage, occasional droughts, frost attack, chronic shortage of cash income, population pressure, poor farming technologies, weak extension services, high labor wastage, poor social and infrastructural situation [21], deforestation, low adoption of agricultural technologies [22], that hamper food security in Ethiopia.

Previous studies that have been done on food security are general. Despite the increasing global concern of ensuring food security, the determinants of food insecurity at the household level in the rural areas have been poorly known, and have not been well investigated [23,24]. In the central part of Southern Ethiopia, smallholder farmers are food insecure whose livelihood is agriculture [23]. Lemo district is one of districts found in the central part of Southern Ethiopia. The mixed farming systems are the tradition by smallholder farmers in the district. The area is vulnerable to undernourishment needing emergency food aid [25]. There are no studies done on investigating determinant of food insecurity in this specific rural area. Therefore, this study was conducted in order to fill this knowledge gap by sharing scientific empirical evidence on food insecurity drivers in rural areas.

2. Materials and methods

2.1. Description of the study area

The study was carried out at Lemo District of Southern Ethiopia (Fig. 1), which lies between 7° 24′ 0″N and 7° 44′ 30″N latitude and 37° 44′ 0″E and 38° 3′ 0″E longitudes. The mean annual rainfall is 800–950 mm and the mean annual temperature is 22 °C. According to the district's agriculture and rural development office, there are various farming systems including livestock rearing, monoculture, agroforestry systems, communal lands, and plantation forests. Nitosols and Leptosols were the major soil type found in the district. The population of the district was estimated about 118,578. Of this 58.104 were males, and 60,474 were females [26].

Fig. 1.

Map of the study rural areas, Lemo district, Southern Ethiopia.

2.2. Research design

2.2.1. Sampling procedure and sample size determination

Multistage sampling technique was used for this study. In the first stage, two Kebeles (Kebele is a lower administration unit in Ethiopia), namely Anabelesa and Digba were selected randomly. In the second stage, enumeration areas were selected randomly. In the third stage, the sample households (HHs) were selected proportionally from each Kebele by using simple random sampling technique. The sample size was determined by Ref. [27]; cited in Israel 1992).

$$\mathrm{n}=\frac{\mathrm{n}1+\mathrm{n}2}{\left(1+\mathrm{N}\left({\mathrm{e}}^2\right)\right)}$$ equation 1

where n = total sample size N = the population size.

e = the acceptable sampling error = 5% = 0.05 at 95% confidence level.

A total sample size used for this study was 383 Households.

Key informants were the elders, administrators, and field expertise who have lived for a long period of time in the villages and who were knowledgeable about food insecurity conditions. The selection of KIs was done by the snowball method [28]. A total of 24 KIs were selected. Key informants were used for stratification of households into food secured households and food insecure households, and data triangulation.

2.2.2. Method of data collection

Both primary and secondary data were collected to identify important variable that may affect household food security. To gather primary data, a semi-structured questionnaire was used to collect quantitative data through household survey from two kebeles, namely, Anabelesa and Digba (Supplementary 1). The household survey included a total of 383 HH heads. The dependent variable in this study was household food security status The quantitative data on the households' food consumption within 7 days were collected. In order to classify the households as food secure or food insecure, by following [29]; 2100 kilocalorie (kcal) a minimum energy requirement an adult leads a heathier and active life in Ethiopia. Households consumed calories above 2100 kcals as food secure and those consumed as food insecure. The independent variables were family size, sex of household heads, age of household heads, education status, off-farm income, landholding size, and agricultural technologies adoption (Supplementary 1). Data were also collected through key informants’ interview for data triangulation.

2.2.3. Method of data analyses

The collected data were analyzed by Statistical Package for Social Sciences version 23.0. The collected data regarding household food consumption within 7 days were converted into kilocalorie by using the food composition kcals per kg of food [30]. The household's daily calorie intake per adult equivalent (AE) was computed using conversion factor (age-sex groups) Storck et al. (1991) cited in Ref. [31]. Finally, the AE per day was divided by household size to get AE per day per person. Descriptive statistics and logistic regression model were used for data analyses. Using a logistic regression model, a response variable can be predicted from a set of continuous, discrete, dichotomous, or a combination of any of these predictor variables.

Let Ynx1 be a dichotomous outcome random variable with categories 1 (food secured household) and 0 (food insecure household). Let Xn x (p + 1) denote the collection of p-predictor variables of Y where $\left[\begin{array}{c}1{\mathrm{x}}_{11}\dots {\mathrm{x}}_{1\mathrm{p}}\\ 1{\mathrm{x}}_{21}{\dots \mathrm{x}}_{2\mathrm{p}}\\ 1{\mathrm{x}}_{\mathrm{n}1}\dots {\mathrm{x}}_{\mathrm{n}\mathrm{p}}\end{array}\right]$ is called regression matrix and without the leading column of 1s is termed as predictor data matrix. We employed pi to represent the probability that Y = 1 and We define 1- pi to the probability that Y = 0.

These probabilities are written in the following form [32].

$$p_i=P\left(Y=1/X_1{X}_2..{.X}_n\right)$$

$${1-p}_i=P\left(Y=0/X_1{X}_2..{.X}_n\right)$$ equation 2

In Equation (2) we employed the model for the natural logarithm of the odds (log odds) to favor Y = 1.

$$\mathit{ln}\frac{P\left(Y=1/X_1{X}_2..{.X}_n\right)}{1-P\left(Y=1/X_1{X}_2..{.X}_n\right)}=\mathit{ln}\frac{p_i}{1-p_i}={\beta }_o+\sum _{j=1}^n{\beta }_j{X}_j\left(1\right)$$ equation 3

Using the inverse of the logit transformation of Equation (3) we arrived at the following:

$$\mathrm{P}\left(\mathrm{Y}=1\backslash /{\mathrm{X}}_1{\mathrm{X}}_2..{.\mathrm{X}}_{\mathrm{n}}\right)=\frac{{\mathrm{e}}^{{\beta }_{\mathrm{o}}+\sum _{\mathrm{j}=1}^{\mathrm{n}}{\beta }_{\mathrm{j}}{\mathrm{X}}_{\mathrm{j}}}}{1+{\mathrm{e}}^{{\beta }_{\mathrm{o}}+\sum _{\mathrm{j}=1}^{\mathrm{n}}{\beta }_{\mathrm{j}}{\mathrm{X}}_{\mathrm{j}}}}=\frac{1}{1+{\mathrm{e}}^{{-(\beta }_{\mathrm{o}}+\sum _{\mathrm{j}=1}^{\mathrm{n}}{\beta }_{\mathrm{j}}{\mathrm{X}}_{\mathrm{j}}}}$$ equation 4

Parameter Estimation: Due to these less stringent underlying assumptions, the maximum likelihood estimation method is suitable for estimating the logistic (logit) model parameters. (Hosmer-Lemeshow, 1989). Hence, in this study the maximum likelihood estimation technique was applied to estimate parameters of the model [32].

3. Ethics statement

Informed consent was obtained from the participants. We, the authors obtained permission to conduct the research from Wachemo University's ethics committee according to the established ethics guidelines of the institution. The researchers have got permission from the Lemo district administration and from the concerned bodies to conduct the study on the topic.

4. Results

4.1. Socio-economic and demographic characteristics of the respondents

Table 1 presented some basic descriptive statistics of the socio-economic and demographic characteristics of the households in the studied rural areas. Out of 383 respondents, 68.4% were food insecure while the rest were food secure. Of 383 households, 362 were male household heads and the rest were female headed households. The results revealed that the respondents who engaged in the off-farm activities had a greater proportion of food security than those who did not participate in off-farm activities in the studied rural areas. Our findings confirmed that the household head that can read and write had a higher proportion of food security as compared with household heads that cannot read and write. According to the results, respondents with larger landholding sizes had higher food security than those with smaller landholding sizes. The proportion of food security increased as the age of the respondents increased in the studied rural areas (Table 1).

Table 1.

Socio-economic and demographic characteristics of the households in the rural area.

Variable	Category	Frequency	Food secure	Food insecure (%)
Sex of household head	Male	362 (94.5)	116 (32)	246 (68)
	Female	21 (5.5)	5 (23.8)	16 (76.2)
Age of household head	< = 50	276 (72.1)	60 (21.7)	216 (78.3)
	>50	107 (27.9)	61 (57)	46 (43)
Off-farm income generating activities	Participate	88 (23)	49 (55.7)	39 (44.3)
	Non-participate	295 (77)	72 (24.4)	223 (75.6)
Education Status	Can read and write	188 (49.1)	77 (39.5)	118 (60.5)
	Cannot read and write	195 (50.9)	44 (23.4)	144 (76.6)
Land size (in hectare)	≤0.5	265 (69.2)	50 (18.9)	215 (81.1)
	>0.5	118 (30.8)	71 (60.2)	47 (39.8)
Technology adoption	Adopter	102 (26.6)	52 (51)	50 (49)
	Non-adopter	281 (73.4)	69 (24.6)	212 (75.4)
Household Size	<6	124 (32.4)	76 (61.3)	48 (38.7)
	> =6	259 (67.6)	45 (17.4)	214 (82.6)
Food security status	Food secure	121 (31.6)
	Food insecure	262 (68.4)

4.2. Determinants of food insecurity in rural households by inferential statistics analysis

The sex of respondents showed the P – values greater than 0.25 by univariate logistic regression analysis, and excluded from the multivariate logistic regression analysis (Table 2). The findings revealed that education status, landholding size in hectare, household size, agricultural technologies adoption, off-farm activities, and age of respondents were significantly contributing to the household food security in the studied rural areas (Table 3).

Table 2.

Results of Univariate logistic regression.

Variable	Coef. (ꞵ)	S.E	Wald	Df	Sig.	Exp (ꞵ)
Sex	−0.411	0.525	0.615	1	0.433	0.663
Age	1.563	0.244	41.116	1	0.000	4.774
Education	0.759	0.226	11.259	1	0.000	2.126
Land size	1.871	0.245	58.341	1	0.000	6.496
Household size	−2.019	0.247	66.940	1	0.000	0.133
Technology	−1.162	0.242	23.093	1	0.000	0.313
Off-farm	−1.359	0.254	28.658	1	0.000	0.257

Table 3.

Determinants of food insecurity in the rural households by Multiple logistic regression.

	ꞵ	S.E.	Wald	d.f	Sig.	Exp (ꞵ)
Constant	−0.398	0.383	1.080	1	0.299	0.672
Education	0.865	0.313	7.667	1	0.006	2.376
Land size	1.982	0.333	35.460	1	0.000	7.260
Household size	−2.423	0.332	53.221	1	0.000	0.089
Technology Adoption	−1.086	0.336	10.458	1	0.001	0.337
Off-farm income	−1.588	0.354	20.086	1	0.000	0.204
Age	1.815	0.338	28.805	1	0.000	6.141

Multiple predictor variables in a logistic regression model were indicated below.

$$\mathrm{Ln}\left(\mathrm{O}\mathrm{D}\mathrm{D}\mathrm{S}\right)=\ln \left(\frac{\hat{Y}}{1-\hat{Y}}\right)=-0.398+0.865\mathrm{e}\mathrm{d}\mathrm{u}+1.982\mathrm{L}\mathrm{s}\mathrm{i}\mathrm{z}\mathrm{e}–2.423\mathrm{h}\mathrm{s}\mathrm{i}\mathrm{z}\mathrm{e}-1.086\mathrm{t}\mathrm{e}\mathrm{c}\mathrm{h}-1.588\mathrm{o}\mathrm{f}\mathrm{f}\mathrm{a}\mathrm{r}\mathrm{m}+1.815\mathrm{a}\mathrm{g}\mathrm{e}$$ equation 5

Odds of household heads who could not read and write were 2.376 times more likely than those who could read and write to experience food insecurity (Table 3). Household with more than half hectare were less likely to experience food insecurity (Table 3). Odds of households who engaged in off-farm income generating activities were 0.204 times less likely to experience food insecurity (Table 3).

Households who did adopt farm technologies were less likely to experience food insecurity than those who did not adopt farm technologies in the rural areas (Table 3).

Odd of respondents with higher-aged were 6.141 times more likely to experience food insecurity than odds of younger-headed households (Table 3). A family with larger number of members were less likely to experience food insecurity than a family with smaller number of members (Table 3).

5. Discussion

In the studied rural households, 68.4% of the sampled populations were food insecure. [11]; reported 71.6% of sampled households were food insecure and 28.4% of households were food secure in Damot Gale district of Wolaita zone, Southern Ethiopia. In contrast, [23]; reported 25.4% of the total population in Woliso district of Western Ethiopia who were vulnerable to food insecurity. In general, the studied rural households were mostly affected by food insecurity. The binary logistic regression model showed that the male headed households were less likely to experience food insecurity than female headed households in the rural areas. This is in agreement with the previous study by Ref. [23]; in southwestern Ethiopia. This discrepancy may be related to cultural norms and behaviors that frequently restrict women's access to resources and participation in food production, preparation, processing, distribution, and marketing activities, which typically impairs women's access to food security and nutrition [10,23].

According to the logistic regression analysis, the sampled households who had the larger cultivating land size in hectare were less likely to be food insecure. The cultivation land is considered as critical agricultural production unit, which determines the food security status of the smallholder farmers under the subsistence agriculture.

This is supported by the previous study done by Ref. [11]; who reported that under subsistence agriculture, the cultivation land size would affect households’ food security.

According to the result, the level of education was negatively associated with food insecurity. The literate households were less likely to be food insecure than uneducated households in the studied rural areas. This could be modernization of agriculture, access of information, and increased creativity by educated household headed ones. This is consistent with another study done by Ref. [15]; who revealed that educated farmers are knowledgeable on how to effectively utilize improved agricultural technologies to improve production. This positively influences the households’ food security status in the rural areas. Another studies supported that an educated household headed had better opportunities to increase their household income, and build assets than uneducated households [10,11].

The logistic regression indicated that the sample households who adopted the agricultural technologies were less risky to be food insecure than non-adopters in the rural areas. This might be the good extension services, access to information and farming experiences. This is in line with the arguments made by Ref. [23]; who reported that new technology adopters were less likely to be food insecure because they can diversify income sources by adopting agricultural technologies.

The ages of the sample households were significant and positively contributed to the households’ food insecurity. Odds ratio indicated that higher-aged household heads were more likely to experience food insecurity than younger-headed households. The reason could be the higher-aged household headers were unproductive while the younger-headed were more productive. This is in agreement by Ref. [16].

In this particular study, it was confirmed that the sampled respondents who had more than five family members were less risky to be food insecure. This might be the larger family size could increase the household income through petty trade, engaging off-farm activities, and wages. This is in agree with [22]; who reported that food insecurity has positive relationship with family size.

Odd of the respondents who engaged off-farm activities were less likely to be food insecure. This could be larger family size engaged in non-farm income activities like petty trade, wages, and livestock production, handcrafting and selling local drinks. This is in agreed with the study done by Ref. [23]. The limitations of this study were excluding determinants of food insecurity in urban areas and factors that may hamper food security in rural areas including policies and strategies.

6. Conclusion

The present study highlighted the important determinants of food insecurity were the educational level of the household headers, landholding size, family size, adoption rate of farm technologies, off-farm activities, age and sex in the studied rural households. The findings of the study revealed that 68.4% of the households were food insecure. Education was an important determinant of households' food security because it was positively related with households’ food security in the rural areas. Educating adults as well as the people who dwell in the rural areas should be mandatory to achieve food security. Agricultural landholding size was a significant factor affecting the food security of rural households. Households who owned low landholding sizes did not overcome food insecurity. Land is shrinking due to population pressure needs agricultural intensification to address food insecurity. Farmers who did adopt improved farm technologies were less likely vulnerable to food insecurity. Access to good extension services and training regarding technologies should be provided to smallholder farmers to improve their productivity. Family size was found to be positively related with the household food security. Off-farm activity was positively correlated with food security. Expanding off-farm activities can play a great role in achieving food security at the grass root levels. Extremely aged household heads were riskier to be food insecure because they were under the unproductive labor category. Family size was found to be positively related with household food security. Understanding the determinants of food insecurity at the grass root level is very essential to achieve food security in the rural community. The results suggest that intervention strategies should be taken by policy-makers, land use planners, decision makers, and development agencies to reduce food insecurity at the grass root levels. Further research should be done on the determinants of household food insecurity in relation to the agro-ecology, policies and strategies of the government.

Declarations

Author contribution statement

Tsidkineh Assefa: Conceived and designed the experiments; Performed the experiments; Analyzed and interpreted the data; Contributed reagents, materials, analysis tools or data; Wrote the paper; Ermias Beyene: Conceived and designed the experiments; Performed the experiments; Analyzed and interpreted the data; Wrote the paper.

Funding statement

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Data availability statement

Data included in article/supp. Material/referenced in article.

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.

Appendix A. Supplementary data

The following is the supplementary data related to this article: