Although various studies have examined the relative influence of farm production diversity and market access on household dietary diversity, few studies have empirically investigated their joint interplay in shaping such diversity. This research addresses this gap based on cross-sectional data from 396 smallholder households in rural Tigray in Northern Ethiopia. In the analysis, a graphical interpretation of the relationship between farm production diversity and market access is followed by a deeper investigation using Poisson estimation methods. We use an alternative measure of market access, households’ frequency of food market visits, rigorously tested for stability, and compare it with alternative measures. Our findings reveal that farm production diversity and market access, jointly and independently, have a positive and significant nonlinear influence on rural smallholder household dietary diversity. The right mix between farm production diversity and smallholder households’ frequency of food market visits increases rural household dietary diversity.
Research Article
The Joint Influence of Farm Production Diversity and Food Market Visits on Smallholder Household Diets in Northern Ethiopia
https://doi.org/10.21203/rs.3.rs-5310265/v1
This work is licensed under a CC BY 4.0 License
Version 1
posted
You are reading this latest preprint version
Household dietary diversity
Market access
Farm production diversity
smallholder households
Frequency of food market visits
Proximity to market
Ethiopia.
Reports indicate that in 2021 around 800 million people were affected by hunger worldwide (FAO et al., 2023). The same reports indicate that global undernourishment remained relatively unchanged in 2022, and that it is on the rise in all subregions of Africa. There is limited information on hunger in Ethiopia, but the latest available data (Bezu, 2018) showed that the number of undernourished people in the country had reached 33 million, which accounts for about 30% of the country’s population. Household food insecurity in Sub-Saharan Africa leads to the consumption of less diverse diets, while the lack of diverse diets, in turn, results in micronutrient malnutrition (Argaw et al., 2021). Ethiopian rural households, especially among smallholder farmers, are characterized by low levels of daily energy consumption, poor nutritional quality and monotonous diets (Dorosh & Minten, 2020). Micronutrient malnutrition, in turn, leaves poor households prone to short-run diseases and long-run cognitive problems (Eckhardt, 2006). In the case of Ethiopia, stunting in children is observed in food-insecure households and those who failed to achieve the minimum dietary diversity (Disha et al., 2013). The development challenge in rural households is twofold: they have limited access to food (Hailu & Amare, 2022; Headey et al., 2019) and the food that they consume is not sufficiently diverse (Dorosh & Minten, 2020; UNDP, 2012). This paper focuses particularly on the second challenge, that of household dietary diversity.
Numerous studies have been undertaken to identify strategies to comprehensively address the challenges related to smallholder household diets. The critical entry point to understanding smallholder household diets is through the major food channels, namely agricultural farm production diversity, and the access to food markets. Although substantial global and local research has been conducted in this domain, the challenge to precisely indicate the pathways shaping household dietary diversity persists, due to mixed findings of various studies (Nandi et al., 2021; Sibhatu & Qaim, 2018). A number of studies have focused on the relative influence of farm production and market access on household dietary diversity, with competing results. Some studies (Jones, 2017; Kumar et al., 2015; Muthini et al., 2020; Romeo et al., 2016; Tesfaye & Tirivayi, 2020) have underlined the greater influence of farm production diversity. Others (Bellon et al., 2020; Bonuedi et al., 2022; Kihiu & Amuakwa-Mensah, 2021; Koppmair et al., 2017; Sibhatu & Qaim, 2017; Usman & Haile, 2022) have emphasized the greater influence of market access. However, there are relatively few empirical studies on the joint interplay between farm production diversity and market access in shaping household dietary diversity. To our knowledge to date, only Zanello et al. (2019) in Afghanistan and Morrissey et al. (2024) in Uganda have examined the interplay between farm production diversity and market access on smallholders’ household dietary diversity, using panel data. This study aims to address the identified gap, capturing the discrete impacts as well as the joint interplay between farm production diversity and market access in defining rural smallholder household diets. This paper introduces an alternative metric for measuring market access compared to most other studies addressing this issue. The common measures of market access, such as distance to nearest market centres, distance to road infrastructures, and so on have been criticized by authors such as Chamberlin & Jayne (2013) for their weakness as measures of farmers’ access to the market. A similar argument against those proxies of market access, states that “these types of proxies for market access may be very poor indicators of the latent quality of a market, particularly the nutritional dimensions of market quality” Headey et al. (2019, p.1312). A systematic review of literature on the subject by Nandi et al. (2021) identified the use of self-reported travel time, distance to the nearest market, time taken, and cost to reach the nearest market in most of the published studies. They criticize this, arguing that “these may not accurately measure the construct of market access and also may not necessarily equate to market participation” Nandi et al. (2021, p.11). Other researchers have recently also questioned the robustness of these traditional measures of market access in explaining the role of markets in modelling household dietary diversity and nutrition (Bonuedi et al., 2022; Matita et al., 2021). As a response to the concerns over the traditional measures of market access, the metric used in this study, the households’ frequency of food market visits, aims to integrate the endogenous and exogenous factors influencing households’ decisions to interact in food markets. Frequency of market visits combines the concepts of market access and market participation in a single measure.
The research is conducted with rural smallholder households in Tigray, Ethiopia. The households were primarily engaged in the production of staple crops, livestock, and limited vegetable and fruit under irrigation, for both consumption and sale. The alternative metric of market access used in this study, the households’ frequency of food market visits, is employed alongside households’ proximity to the market and length of households’ absence from food markets, to compare and test the validity and robustness of the metric. The latter measure explores the implications when households refrain from purchasing foods for longer periods. We measure this through the number of days since the household’s last food purchase and refer to it as household’s absence from food markets. We hypothesize a significant positive contribution of the discrete and joint interaction of farm production diversity and frequency of food market visits on household dietary diversity.
Our empirical data is analysed based on a cross-sectional food consumption survey from rural smallholder households in Tigray. Following a graphical analysis of the relationships to be investigated in the survey data, the Poisson estimation model is used to analyse the collected data.
Section 2 introduces the study area, the techniques of data collection, the information collected, and the method used to analyse the data. Section 3 then discusses the descriptive statistics and econometric findings followed by the conclusions in Section 4. Relevant tables are included in the Appendix.
2.1. Description of the Gergera Watershed in Tigray, Ethiopia
This research was conducted in the Gergera watershed, located in the rural administrative area of Hayelom Tabia in the regional state of Tigray, northern Ethiopia (See Figure 1). Geographically, the Gergera watershed is situated in the Eastern zone of Tigray, at a latitude of between 13.043 and 13.046N and a longitude of 39.040 to 39.045E. The watershed covers a land area of 2,302 hectares. The average daily temperature ranges from 15°C to 25°C (ICRAF, 2018). According to the local agroecological classification, the Gergera watershed lies within highlands (dega) and mid-land (woyna dega) areas, with an altitude ranging between 2,066 and 2,505 meters above sea level. The weather is characterized as semi-arid, with a bimodal annual rainfall regime between 450mm and 600mm. The main rainfall occurs in the summer season from mid-June to early September, and this is the main season for crop cultivation. There is another short rainy season between March and April which the farmers usually use for cropland preparation.
Morphologically, the Gergera watershed is characterized by an undulating terrain with rugged topographic features. The soil texture is mainly silt-loam and loam with neutral PH and very little alkaline (Woldewahid et al., 2012). Mixed crop and livestock are the dominant farming practice in the watershed. The main crops grown are cereals, such as barley, wheat, teff, maize, and sorghum; and pulses, including beans, peas, and lentils (Deribe, 2008).There are some irrigation activities on the lower areas of the valley around the outlet of the watershed. Vegetables, such as tomato, onion, cabbage, and spices, and fruits such as guava, avocado, papaya, and orange are some of the major products of the irrigable fields (Woldewahid et al., 2011). The livestock in the area includes cattle, sheep, goats, donkeys, horses, and chickens (Gessesse, 2016).The flat areas and enclosures on the hillsides are usually reserved for fodder grass collection and beekeeping (Meaza, 2010).
2.2. Data Sources and Sampling
Primary data was collected using a cross-sectional survey conducted with rural smallholder households from the four sub-districts (kushets[1]) of the Gergera watershed, namely Geter Haiki Mesihal, Damaino, Gergera, and Degaabur. We applied a stratified random sampling technique to select a sample from the population. A total of 396 households were selected in proportion to the population size in the four sub-districts. Systematic random sampling was used to pick individual households from the list of residents provided by the local administrative office. The data was collected in the period between September and October 2020.
The actual data collection was executed by experienced enumerators following a comprehensive training on the translated questionnaire and undergoing pilot surveys to evaluate the suitability and efficiency of the survey.
Data was collected using a four-part questionnaire. The first part deals with the households’ demographic and socioeconomic characteristics; the second focuses on household food consumption data; the third on the household farm production diversity; and the fourth on households’ market interaction. Crop and livestock production data for one agricultural year was captured, covering the period from September 2019 to August 2020.
Our statistical analysis integrates the findings obtained from the simple statistical analysis and the econometric estimations. This approach not only increases the dependability of the results but also highlights the robustness of our research framework.
2.3. Measurement of Key Variables
2.3.1. Measuring Household Dietary Diversity
The household dietary diversity score (hereafter HDDS) is used to measure the diversity of household access to food, using a 24-hours recall period (Swindale & Bilinsky, 2006; FAO, 2007; Sibhatu & Qaim, 2018; Admasu et al., 2011). According to the Swindale & Bilinsky (2006) Food Access Indicator Guide, food items can be classified into 12 food groups, based on their nutritional contribution, to compute the HDDS. The 12 food group classifications are as follows: 1. Cereals; 2. White tubers and roots; 3. Vegetables; 4. Fruits; 5. Meat; 6. Eggs; 7. Fish and other seafood; 8. Legumes, nuts, and seeds; 9. Milk and milk products; 10. Oils and fats; 11. Sweets and honey and, 12. Spices, condiments, and beverages (Swindale & Bilinsky, 2006). From this list of food groups, households were asked to identify any food item they consumed in the previous 24 hours. The enumerators recorded ‘1’ if the household had consumed an item from a food group or ‘0’ otherwise. The sum of the number of food groups consumed in the 24 hours gives the total value of the HDDS for the household, which may range from a minimum of zero, if no food at all was consumed in the 24-hours, to a maximum of twelve, if the household consumed something from all 12 food groups. Only one point is recorded if a household consumes from a particular food group within 24 hours, regardless of the quantity or the frequency of the consumption within that time, or the number of people who consumed it. For example, a household would likely consume injera[2] (the local staple food) for up to four meals in a day. According to the HDDS, however, the repeated consumption of injera is counted as one (being part of the cereal food group). Following the HDDS methodology, food consumed outside the home, such as in markets, public events, or at neighbours, is not counted as part of the household diet. The computation of HDDS can be presented as follows:
2.3.2. Measuring Farm Production Diversity
Different methods are available for measuring farm production diversity. In this study, we have adopted the production diversity score (hereafter, PDS) to measure the farm production diversity, calculated in terms of the number of food groups produced by households in the previous agricultural year (September 2019 to August 2020). The PDS classifies crops and livestock into 12 food groups, effectively the same as for the HDDS discussed above (Chegere & Stage, 2020; Koppmair et al., 2017; Sibhatu & Qaim, 2018). We have constructed our data for the farm production diversity score based on the 12 food groups that may be produced at a smallholder household level. The food groups include: 1. Cereals; 2. White tubers and roots; 3. Vegetables; 4. Fruits; 5. Meat; 6. Eggs; 7. Fish and other seafood; 8. Legumes, nuts, and seeds; 9. Milk and milk products; 10. Oils and fats; 11. Sweets and honey and, 12. Spices, condiments, and beverages (Swindale & Bilinsky, 2006).
The application of the PDS has some limitations arising from possible multiple counting of certain food groups, particularly animal products (Berti, 2015; Kissoly et al., 2020). A PDS value of two is recorded for a household keeping chickens, because chickens may produce two different food groups, meat, and eggs. Similarly, a cow can be a source of three different food groups: meat, milk, and butter (oil food group). On the other hand, an ox is given a score of 1, whether it is consumed in the given year or not. PDS is therefore, seen as a count of potential food sources, rather than a record of actual consumption from own-produced sources.
Muthini et al. (2020a) have excluded non-food crops from the PDS count. Although no such crops are produced in the study area, we follow this example by excluding animals such as donkeys, horses, mules, and camels, since these are working animals and are not considered foods in the local culture, thereby falling outside of the standard definition of PDS as a count of food groups.
2.3.3. Measuring Households’ Access to Market
Previous literature has employed proxy measures, such as distance to the market, distance to roads or other public infrastructures, cost of transportation to the market, and so on, to assess market access (Hoddinott et al., 2015; Jones et al., 2014; Sibhatu et al., 2015; Koppmair et al., 2017; Muthini et al., 2020; Zanello et al., 2019). However, those proxy measures of market access have been criticized for their weakness in explaining the influence of markets in shaping diets and nutrition (Bonuedi et al., 2022; Chamberlin & Jayne, 2013; Headey et al., 2019; Matita et al., 2021; Nandi et al., 2021). These proxy measures of market access are mostly exogenous (Usman & Haile, 2022), neglecting the households’ social, economic, cultural, and other factors that can influence a decision to engage in market transactions. Headey et al. (2019) for instance has criticized those measures and employed a combination of the market diversity, the numbers of traders selling foods, and other variables instead of the traditional market access measures. Those measures have, however, focused on the quality of the market and gives little attention to the household ability to take advantage of the existing market. This paper, therefore, uses an alternative metric of market access which is based on household decisions that implicitly integrates the internal and external factors influencing their market engagement.
The preferred measure of market access in this paper is, therefore, the households’ frequency of food market visits in 30 days with the aim to purchase food items or in short, the Frequency of food Market Visits (FMV). The frequency of food market visits (FMV) is a comprehensive measure of smallholder farmers’ access to the market, reflecting not only the physical proximity to the market, but also nuanced dimensions of households' preferences, willingness, and ability to participate in market transactions. In support of this measure, we also use the length of days households abstain from food markets since their last visit, as a second alternative measure of market access. Examination of the length of absence from the food markets facilitates the evaluation of how households’ level of dietary diversity is impacted if they refrain from purchasing foods for an extended period. Our measure of market access is preferred compared to the distance in kilometres approaches because (i) it considers market access as an outcome of households’ decision to transact in a market, (ii) the decision made is related to food purchases which has a direct influence on the households’ dietary diversity, (iii) it is flexible (compared to the static measure of distance) in response to various factors including social, economic, cultural, and political factors, affecting households decision to engage in market transactions in a specific period. We collected data on households’ frequency of food market visits in 30 days, a period that can be well remembered by the household. We have selected 30 days to give space for a range of variations among household market engagements. To compare the results and check the robustness of the frequency of food market visits (FMV), we included two alternative metrics of market access in this paper. The first is the households’ proximity to the food markets (PM), measured in walking minutes, an established metric of market access (Bergau et al., 2022; Koppmair et al., 2017). We also added absence from food markets (AM) which indicates the number of days elapsed since the household's last food market visit. This measure of market access captures the difference in household dietary diversity on market days and non-market days. The approach gives a direct observation of market’s influence on HDD.
2.4. Data Analysis
2.4.1. Poisson Regression Equation
The Poisson regression equation is based on the count outcomes of the number of food groups consumed by households in 24 hours. The count outcomes of food group consumption by households are assumed to have the characteristics of a Poisson probability distribution function. This distribution function is revealed by the occurrence of non-negative discrete outcomes at a fixed interval of time, where each event is independent of one another. The Poisson regression equation is used to model the probability that a household consumes a fixed number of food groups in 24 hours, given their household and socioeconomic characteristics. Mathematically, the Poisson regression equation takes the form:
In the regression models, the dependent variable HDDSi represents the household i’s extent of dietary diversity in the 24 hours. The key explanatory variables are the farm production diversity score PDSi and market access. Market access is alternatively explained by the household ’s frequency of food market visits, FMVi. Absence from the market (length of days since last food market visit) AMi and household i’s Proximity to the market centre PMi. We hypothesise that both PDSi and FMVi have a positive association with HDDS, while AMi and PMi have negative correlations with HDDSi.
SEi represents the vector of the socioeconomic characteristics of the household which includes age of the head, sex of the head, family size, level of education, and marital status of the head, owned rainfed farmland holding size, irrigable land holding size, and household residential sub-district (kushet).
[1] Kushet is the lowest or seventh governmental administrative level in Ethiopian rural governance structure next to Tabia.
[2] Injera is a pancake made of teff, a cereal crop and Ethiopia’s staple crop. Almost every meal in Ethiopian family has injera made of teff, or sometimes made of wheat, maize, sorghum or a mixture of those cereal crops.
3.1. Descriptive Statistics
The objective of this study was to evaluate the discrete as well as the mutual interplay between farm production diversity and market access in influencing household dietary diversity. We have used primary data collected through a cross-sectional household survey in Gergera watershed, in eastern Tigray, northern Ethiopia. Data was collected from four sub-districts/kushets of the Gergera watershed. The detailed report from the descriptive statistics is presented in Table 1, below. The 120 households were selected from the subdistrict Damaino, 116 from Geter Haiki Mesihal, 84 from Degaabur, and 76 from Gergera. The gender-based proportions of the surveyed households are 290 male-headed households to 106 female heads. The mean age of the heads was 54.05 years, and the mean household size was 5.11 members. The average household dietary diversity score (HDDS) in the study area was 5.86 food groups per day per household (Table 1), while the mean farm production diversity was 3.48 food groups per year. The farm production diversity score shows a difference across genders of the household heads: male-headed households’ mean annual farm production diversity (PDS) was 3.70. Female-headed households’ mean annual farm production diversity level was 2.90 food groups. The mean monthly frequency of food market visits for all households was 3.03. Male-headed households made more frequent food market visits, with a mean of 3.13 times in 30 days, compared to female-headed households, with a mean of 2.75 times in 30 days. Households’ frequency of food market visits across subdistricts reveals a mixed association with households’ proximity to the market centre.
Subdistrict Geter Haiki Mesihal has the shortest mean walking time to the food market at 27.17 minutes, with a mean frequency of food market visits of 3.63; this is followed by Gergera, with a mean walking distance of 42.37 minutes and a slightly higher rate of 3.83 food market visits per month (Table 1). Residents of Damaino and Degaabur, the third and the fourth remotest subdistricts (a mean travel time of 82.42 and 149.88 walking minutes, respectively) made considerably less visits to the market, at 2.84 and 2.13 visits, respectively (Table 1).
Examining the third market access measurement variable, (Table 1), the results show that absence from the market demonstrates mixed association with proximity to the market. Residents of Dega’abur, the remotest subdistrict, demonstrate a mean of 5.11 days, while those in Geter Haiki-Mesihal, the closest to the market, had a mean of 4.77 days (Table 1). Those at an intermediate distance did not form a pattern: residents of Damaino reported the longest absence from food markets, with a mean of 7.76 days, while those in Gergera had the shortest absence from the market, at 3.77 days on average. Gender-wise, female-headed households experience longer absence from food purchases, with an average of 7.20 days, compared to male-headed households, with a mean of 6.26 days of absence (Table 1).
Table 1: Summary statistics for selected socioeconomic characteristics.
|
HDDS |
PDS |
Freq. food market visits (in 30 days) |
days of absence from market |
Proximity to market (In walking minutes) |
Age |
Family size |
Education |
Rainfed Land size (In Tsimad) |
Irrigable land size (In Tsimad) |
|
|
GENDER |
|
|||||||||
|
Female: N |
106 |
106 |
106 |
106 |
106 |
106 |
106 |
106 |
106 |
106 |
|
Mean |
5.53 |
2.90 |
2.75 |
7.20 |
75.75 |
51.58 |
3.42 |
0.37 |
0.77 |
0.22 |
|
|
|
|||||||||
|
Male: N |
290 |
290 |
290 |
290 |
290 |
290 |
290 |
290 |
290 |
290 |
|
Mean |
5.99 |
3.70 |
3.13 |
6.26 |
71.8 |
54.96 |
5.73 |
1.54 |
1.33 |
0.32 |
|
MARITAL STATUS |
|
|
||||||||
|
Divorced: N |
46 |
46 |
46 |
46 |
46 |
46 |
46 |
46 |
46 |
46 |
|
Mean |
5.50 |
2.96 |
2.98 |
7.02 |
62.93 |
50.07 |
2.87 |
0.37 |
0.74 |
0.2 |
|
|
|
|||||||||
|
Married: N |
287 |
287 |
287 |
287 |
287 |
287 |
287 |
287 |
287 |
287 |
|
Mean |
6.00 |
3.71 |
3.15 |
6.12 |
72.46 |
54.3 |
5.81 |
1.6 |
1.32 |
0.33 |
|
|
|
|||||||||
|
Single: N |
2 |
2 |
2 |
2 |
2 |
2 |
2 |
2 |
2 |
2 |
|
Mean |
6.50 |
2 |
1.50 |
9.00 |
120 |
25 |
3 |
0 |
0.5 |
0.12 |
|
|
|
|||||||||
|
Widowed: N |
61 |
61 |
61 |
61 |
61 |
61 |
61 |
61 |
61 |
61 |
|
Mean |
5.48 |
2.87 |
2.59 |
7.89 |
80.66 |
56.85 |
3.57 |
0.16 |
0.86 |
0.23 |
|
SUB-DISTRICTS |
|
|
||||||||
|
Geter Haki Mesihal: N |
116 |
116 |
116 |
116 |
116 |
116 |
116 |
116 |
116 |
116 |
|
Mean |
6.06 |
3.69 |
3.63 |
4.77 |
27.17 |
56.49 |
5.26 |
1.66 |
1.23 |
0.3 |
|
|
|
|||||||||
|
Damaino N: |
120 |
120 |
120 |
120 |
120 |
120 |
120 |
120 |
120 |
120 |
|
Mean |
5.82 |
3.42 |
2.84 |
7.76 |
82.42 |
49.76 |
5.08 |
1.11 |
1.26 |
0.46 |
|
|
|
|||||||||
|
Gergera N: |
76 |
76 |
76 |
76 |
76 |
76 |
76 |
76 |
76 |
76 |
|
Mean |
6.03 |
3.82 |
3.84 |
3.77 |
42.37 |
57.92 |
5.24 |
1.49 |
1.03 |
0.35 |
|
|
|
|||||||||
|
Degaabur N |
84 |
84 |
84 |
84 |
84 |
84 |
84 |
84 |
84 |
84 |
|
Mean |
5.51 |
3 |
2.13 |
5.11 |
149.88 |
53.32 |
4.82 |
0.56 |
1.11 |
0.00 |
|
TOTAL: N |
396 |
396 |
396 |
396 |
396 |
396 |
396 |
396 |
396 |
396 |
|
Mean |
5.86 |
3.48 |
3.03 |
6.49 |
72.86 |
54.05 |
5.11 |
1.23 |
1.18 |
0.3 |
Source: authors’ own computation from survey data collected in Sep. – Oct. 2020.
Gender-wise evaluation of the summary statistics in Table 1 demonstrates that male household heads have a higher average age, compared to female- heads, at 54.96 and 51.58 years, respectively; higher family size, at 5.73 to 3.42 persons; more years of schoolings, at 1.54 to 0.37; bigger rainfed farmland size, 1.33 to 0.77 hectares; and bigger irrigable landholding size, 0.32 to 0.22 hectares. Overall, these results show a dominance of male heads of households in all the parameters examined in Table 1.
3.2. Simple Statistical Analysis
3.2.1. Independent Relationship between Farm Production Diversity and Household Dietary Diversity
In Figure 2 mean Farm Production Diversity Score (PDS) is graphed on the X-axis, and the average values of household dietary diversity score (HDDS) on the Y-axis. The graph shows the independent relationship between the two variables. The visualization indicates that there is a positive non-linear relationship between farm production diversity and household diversity which can be divided into two stages of returns. The linear and LOWESS fit confirm that there is a positive return in the early stages of increasing farm production diversity, followed by diminishing returns to each additional unit of farm production diversity in the intermediate and later stages. A closer look at Figure 2 depicts that at the initial stage, additional food group production leads to a greater increase in HDDS.
Figure 2: Discrete Relationship between Farm Production Diversity and Household Dietary Diversity
At the intermediate stage, the rate of increase in HDDS diminishes for every additional PDS level; and gets flatter at later stages about 5 PDS levels and beyond.
3.2.2. Independent Relationship between Market Access and Household Dietary Diversity
In Figure 3 (A-C) a set of graphs are presented to examine the independent influence of market access on household dietary diversity. The three different graphs are used to examine the results and compare if there are significant differences when different measures of market access are used.
Figure 3A shows a distinct relationship between households’ frequency of food market visits and their level of household dietary diversity. The household dietary diversity is treated as a dependent variable, on the Y-axis, and the frequency of food market visits (FMV), on the X-axis. The graph shows that every additional market visit positively contributes to the household diet. Contrary to the pattern we observed on the PDS, the FMV shows a smaller rate of additions to the HDDS at the initial stages, and the rate of contribution increases with more market visits. A closer observation of the graph suggests that a certain level of market visits can achieve the maximum HDDS, beyond which each marginal visit brings little or no addition, but without turning negative Figure 3B depicts that recent market visits bring diverse foods to rural smallholder households. Market-day diets are more diverse than other days. Household dietary diversity diminishes as the number of days since a market visit increases. This is a factor for most of the surveyed households who live in rural areas, where there is limited electricity coverage and hence limited food preservation technologies. Households may purchase perishable foods that need to be consumed fresh, such as meat, milk, vegetables, and fruits.
Another interesting phenomenon revealed in Figure 3B is that when households attend food markets regularly (when their absence rate is small), they enjoy higher HDDS levels. However, a longer absence rate leads to lower but not zero HDDS: other food sources (for most households, primarily own-production) fill the gap in the absence of food purchases.
In Figure 3C we have presented another measure of market access, proximity to the market measured in walking minutes. Average proximity to the market is on the X-axis and average HDDS on the Y-axis. The graph indicates an inverse relationship between HDDS and proximity to the market: longer walking minutes adversely impact HDDS. Following the LOWESS fit, Figure 3C has also its own unique features. Close proximity to market is associated with higher HDDS but there is little difference among households living within 50 walking minutes radius. The downward trend in the middle of the graph shows that the remoter residents suffer from lower household dietary diversity levels. The horizontal line at the end of the graph (bottom right, which is visible in LOWESS fit) signals that after a certain remoter distance, the negative influence of remoteness on HDDS will be negligible. Accordingly, little or no change in HDDS is expected due to further increased distances if the households reside beyond 95 walking minutes from the food markets, in our study area.
3.2.3. Joint Relationship between Market Access and Farm Production Diversity
The joint interplay between farm production diversity and market access in influencing household dietary diversity is graphically examined in Figure 4A-C which graphs PDS against the three market access measures used in this study. Figure 4A shows that farm production diversity encourages households to make more food market visits but only up to a certain optimal point. As depicted in Figure 4A, the frequency of food market visits sharply increases with an increase in farm production diversity level at the initial stages with a lower rate of increase in visits at later stages.
Figure 4B graphs PDS (x-axis) against days of absence from markets (y-axis), and not surprisingly shows an inverse relationship compared with Figure 4A. This figure shows that more regular market attendance is associated with higher farm production diversity levels. However, there is little association beyond PDS levels of 5 or above. The final comparison, Figure 4C, shows the association between households’ proximity to markets and their level of farm production diversity. Farm production diversity is higher in households closer to the markets than in remote areas. The relationship implies that proximity to food markets encourages households to pursue more diverse farm production, while remoteness discourages farm production diversity.
Overall, we find that market access has a strong positive influence on farm production diversity. This result is contrary to the expectation that households in remote areas tend to diversify to compensate for inability to make food purchases due to limited access to the market.
3.3. Econometrics Results
The Poisson estimation evaluates the discrete and joint influence of farm production diversity and market access on household dietary diversity. Results are shown in Table 2. We have established two sets of models: the first set (Columns 1 to 4) is used to evaluate the discrete influence of farm production diversity (PDS) and the three market access measures (frequency of food market visits (FMV), absence from market (AM), and proximity to market (PM)) on household dietary diversity (HDDS). The second set of models in Table 2 (Columns 5 to 10) shows the joint influence of farm production diversity and the three market access measures on household dietary diversity.
Later in Table 3, HDDS is explained by the combination of the PDS, the three market access measures, and other socioeconomic factors. The results in Table 2 and Table 3 are presented in terms of marginal effects of the coefficients at the mean, from Poisson regressions.
3.3.1. The Independent Influence of Farm Production Diversity and Market Access on HDDS
The output in Table 2 Column 1 shows that the discrete influence of farm production diversity on household dietary diversity is positive and significant. This result is similar to previous findings based on various approaches such as Kumar et al. (2015); Jones et al. (2014); Romeo et al. (2016); Tesfaye and Tirivayi (2020); Muthini et al. (2020a). An additional unit of farm production diversity increases HDDS by 0.285 food groups, significant at a 1% level.
Independent evaluation of the three market access measures on HDDS is examined in Columns 2 to 4. All the three market access measures show a highly significant association with HDDS at a 1% level. However, the magnitude and direction of their influence is different. The frequency of food market visits has positive and higher magnitude of influence, compared to the remaining two market access measures. This finding is similar to the results obtained using panel data from Uganda by Morrissey et al. (2024). Every additional food market visit increase HDDS by an average of 0.452 food groups, (Column 2). Households’ absence from the market (AM) and proximity to market (PM) show negative influence with magnitudes of -0.011 for AM, (Column 3) and -0.004 for PM, (Column 4). The results indicate that household dietary diversity declines by an average of 0.11 food groups, if the household is absent from the market for 10 days.
Table 2: Poisson estimation of the influence of farm production diversity (PDS) and market access on HDDS
|
Household Dietary Diversity Score/ HDDS |
Column (1) |
Column (2) |
Column (3) |
Column (4) |
Column (5) |
Column (6) |
Column (7) |
Column (8) |
Column (9) |
Column (10) |
|
Production Diversity Score/PDS |
0.285*** |
|
|
|
0.205*** |
0.223*** |
0.271*** |
|
|
|
|
(0.037) |
(0.038) |
(0.038) |
(0.037) |
|||||||
|
Frequency of food market visits |
0.452*** |
0.338*** |
||||||||
|
(0.058) |
(0.059) |
|||||||||
|
Absence from market |
-0.011** |
-0.088*** |
||||||||
|
(In number of days) |
(0.004) |
(0.015) |
||||||||
|
Proximity to Market |
-0.004*** |
-0.003*** |
||||||||
|
(Walking minutes) |
(0.001) |
(0.001) |
||||||||
|
Frequency of visits X PDS |
0.070*** |
|||||||||
|
(Interaction) |
(0.007) |
|||||||||
|
Absence X PDS |
-0.012** |
|||||||||
|
(Interaction) |
(0.004) |
|||||||||
|
Proximity X PDS |
0.000 |
|||||||||
|
(Interaction) |
|
|
|
|
|
|
|
|
|
(0.000) |
|
Observations |
396 |
396 |
396 |
396 |
396 |
396 |
396 |
396 |
396 |
396 |
Table 2 presents the marginal effects after Poisson estimation. The dependent variable is the household dietary diversity score, based on the 12 food groups. The models were estimated using Poisson model. The coefficient estimates are shown in column (1) to (4), robust standard errors in brackets. * p<0.05, ** p<0.01, *** p<0.001.
Source: authors’ own computation from survey data collected in Sep. – Oct. 2020.
Likewise, every additional 10 walking minutes farther from the market leads to an average of 0.04 food groups lower levels of HDDS.
3.3.2. The Independent Influence of Farm Production Diversity and Market Access on HDDS
The other key objective of this study is to examine the joint interplay between farm production diversity and market access on rural smallholder household dietary diversity. The output for the joint treatment of the variables is presented in Table 2, (Columns 5 to 10). The joint influences are estimated in two categories of models. The first category (Columns 5 to 7) evaluates the PDS pairing with each of the three market access measures. The second category (Columns 8 to 10) examines the interaction (multiplication) between the PDS and each of the three market access measures.
Evaluation of the joint interplay in Columns 5 to 7 of Table 2 reveals a positive but lower magnitude of influence, 0.205 for farm production diversity and 0.333 for the frequency of food market visits, compared to the results in Columns 1 and 2 (treated independently). However, the result confirms a combined positive influence of the PDS and the FMV on HDDS.
The joint estimation of farm production diversity and absence from the market in Table 2 (Column 6) indicates a lower magnitude of the PDS, 0.223, and a higher negative influence of absence from the market, -0.088, on HDDS, significant at 1% level. Conversely, the household’s proximity to the market shows a -0.003 influence on HDDS. Bergau et al. (2022) have also found a -0.001 magnitude of farmers’ travel time on HDDS in Ethiopia, which is close to this finding. The magnitude of the impact of PDS on HDDS in Table 2 has reduced from 0.285 (Column 1) to 0.271 (Column 7). This is possibly due to the negative effect of remoteness on PDS. The result is also supported by the visualization presented in Figure 4C above. One important point is that market access’ magnitude of influence measured in the frequency of food market visits is higher than in travel time as in Bergau et al. (2022). This indicates that how market access is measured affects its magnitude of influence (Muthini et al., 2020b).
Table 2 (Columns 8 to 10) depicts the interaction (multiplication) between the PDS and the three market access measures in examining their joint interplay. In Table 2 (Column 8), the magnitude of the interaction between PDS and the frequency of food market visits becomes 0.070, which is positive and significant at the 1% level. This value suggests a positive interaction between the two. The high magnitude of the coefficient of the FMV in different models, its partial derivative on HDDS, and using the effect plotting technique, the result indicates that the joint effect is altered by the households’ frequency of food market visits. This implies that keeping all other factors constant, where PDS and FMV are in operation, the frequency of the food market visits determines the level of household dietary diversity. The results reaffirm that the joint interplay between the frequency of food market visits and farm production diversity enhances smallholder household dietary diversity.
The interaction between PDS and absence from food markets (AM) is found be negative (-0.012), and significant at a 1% level, (Column 9). This suggests that these two variables have an inverse relationship. However, when both are simultaneously in operation, the household’s length of absence from food markets determines the HDDS. This means a household with a relatively high farm production diversity is likely to have a low level of dietary diversity due to its low food market attendance rate. Table 2 (Column 10) also shows the interaction between PDS and household proximity to the market. The relationship is insignificant. This insignificance might be associated with the lower consistency and weak predictive capability of proximity to the market as a measure of market access. Similar arguments have been made by Chegere & Stage (2020), and Usman & Haile (2022).
Overall, examination of the joint interplay between PDS and FMV reveals that frequent food market visits combined with greater farm production diversity, reinforce one another in improving household dietary diversity. A recent study conducted in Uganda comes to a similar conclusion, stating that “there are benefits to smallholders that come from a combination of crop diversity and market engagement” Morrissey et al. (2024, p.11).
Our study highlights the importance of considering the frequency of food market visits as an alternative measure of market access in understanding household dietary diversity. The results provide insights into the complex interplay of farm production diversity and market access in influencing household dietary diversity. This intuition can inform policies and interventions designed to improve market access and household dietary diversity in rural areas.
3.3.3. Estimation of the PDS and Market Access Alongside Control Variables
After estimating the influence of the PDS and FMV on HDDS separately, we regressed them together with smallholder household and socioeconomic characteristics as presented in Table 3. This can identify other influencing factors that might have been omitted in previous estimations in Tables 1 and 2. Moreover, it helps to test the robustness and stability of the key explanatory variables across different models. There are 6 models estimated using the Poisson regression and their marginal effects are indicated in Table 3. Columns 1 to 3 indicate the joint interplay among the PDS, the three alternative market access measures, and other selected socioeconomic variables. Columns 4 - 6 take the interaction of the PDS with the three market access measures and the control variables.
Table 3: Marginal effects after Poisson estimation:
The influence of farm production diversity, market access, and control variables on household dietary diversity.
|
Household dietary diversity score (HDDS) |
-1- |
-2- |
-3- |
-4- |
-5- |
-6- |
|
|
HDDS |
HDDS |
HDDS |
HDDS |
HDDS |
HDDS |
|
Production diversity Score/PDS |
0.208*** |
0.198*** |
0.255*** |
|||
|
(0.037) |
(0.037) |
(0.038) |
||||
|
Frequency of food market visits |
0.332*** |
|||||
|
(0.068) |
||||||
|
Absence from market |
|
-0.095*** |
|
|
|
|
|
(In number of days) |
|
(-0.017) |
|
|
|
|
|
Proximity to market |
-0.003 |
|||||
|
(Walking minutes) |
(0.002) |
|||||
|
Frequency X PDS |
0.066*** |
|||||
|
(interaction) |
(0.008) |
|||||
|
Absence X PDS |
|
|
|
|
-0.010*** |
|
|
(Interaction) |
|
|
|
|
(0.005) |
|
|
Proximity X PDS |
0.001*** |
|||||
|
(interaction) |
(0.000) |
|||||
|
Marital Status |
||||||
|
Divorced |
0.000 |
0.000 |
0.000 |
0.000 |
0.000 |
0.000 |
|
(.) |
(.) |
(.) |
(.) |
(.) |
(.) |
|
|
Married |
0.184 |
-0.003 |
0.238 |
0.245 |
0.264 |
0.162 |
|
(0.49) |
(0.377) |
(0.468) |
(0.469) |
(0.437) |
(0.463) |
|
|
Single |
1.387*** |
1.367*** |
1.485*** |
1.434*** |
1.765*** |
1.211** |
|
(0.386) |
(0.324) |
(0.428) |
(0.384) |
(0.373) |
(0.523) |
|
|
Widowed |
-0.013 |
0.036 |
0.102 |
0.032 |
0.12 |
-0.057 |
|
(0.27) |
(0.263) |
(0.265) |
(0.263) |
(0.265) |
(0.271) |
|
|
Sub-district |
||||||
|
Geter Haiki Mesihal |
0.000 |
0.000 |
0.000 |
0.000 |
0.000 |
0.000 |
|
(.) |
(.) |
(.) |
(.) |
(.) |
(.) |
|
|
Damaino |
-0.621*** |
-0.001 |
-0.036 |
-0.083 |
0.031 |
-0.246 |
|
(0.183) |
(0.142) |
(0.15) |
(0.196) |
(0.156) |
(0.16) |
|
|
Gergera |
-0.128 |
-0.114 |
0.018 |
-0.015 |
0.013 |
-0.041 |
|
(0.198) |
(0.173) |
(0.18) |
(0.184) |
(0.179) |
(0.189) |
|
|
Degaabur |
-0.819*** |
0.164 |
0.038 |
0.167 |
0.189 |
-0.217 |
|
(0.223) |
(0.173) |
(10.175) |
(0.299) |
(0.185) |
(0.187) |
|
|
Gender |
||||||
|
Female |
0.000 |
0.000 |
0.000 |
0.000 |
0.000 |
0.000 |
|
(.) |
(.) |
(.) |
(.) |
(.) |
(.) |
|
|
Male |
0.018 |
0.109 |
-0.018 |
-0.066 |
-0.048 |
0.007 |
|
(0.416) |
(0.300) |
(0.398) |
(0.396) |
(0.369) |
(0.387) |
|
|
Education level |
0.044* |
0.031 |
0.040* |
0.042* |
0.038* |
0.041 |
|
(0.024) |
(0.022) |
(0.023) |
(0.023) |
(0.022) |
(0.025) |
|
|
Age |
-0.004 |
-0.003 |
-0.003 |
-0.005 |
-0.003 |
-0.001 |
|
(0.005) |
(0.005) |
(0.005) |
(0.005) |
(0.005) |
(0.005) |
|
|
Family size |
0.033 |
0.024 |
0.021 |
0.021 |
0.011 |
0.060* |
|
(0.033) |
(0.032) |
(0.033) |
(0.033) |
(0.033) |
(0.033) |
|
|
Rainfed farmland holding size |
0.047 |
-0.003 |
-0.056 |
-0.016 |
-0.064 |
0.108 |
|
(In tsimad= 1/4 hectare) |
(0.062) |
(0.057) |
(0.060) |
(0.061) |
(0.060) |
(0.063) |
|
Irrigable landholding size |
0.595*** |
0.495*** |
0.320** |
0.446*** |
0.275* |
0.568* |
|
(In tsimad= 1/4 hectare) |
(0.150) |
(0.136) |
(0.137) |
(0.138) |
(0.140) |
(0.151) |
|
Observations |
396 |
396 |
396 |
396 |
396 |
396 |
Table 3 presents the marginal effects after Poisson estimation. The dependent variable is the household dietary diversity score, based on the 12 food groups. The models were estimated using Poisson model. The coefficient estimates are shown in column (1) to (4), robust standard errors in brackets. * p<0.05, ** p<0.01, *** p<0.001.
Source: authors’ own computation from survey data collected in Sep. – Oct. 2020.
The results in Table 3 show that PDS regressed with selected control variables positively and significantly influences HDDS, ranging from 0.198 to 0.225 food groups. There are small changes in coefficient values compared with Table 2 which could be associated with the compensating effects of the included control variables: nonetheless, the direction and level of significance observed in Table 2 is sustained. Similarly, the frequency of food market visits with an estimated coefficient of influence of 0.332, and that of the absence from the market, -0.095, show only minor differences compared to their respective magnitude in Table 2, and the direction and level of significance are still maintained. However, although proximity to the market has the same magnitude of influence as was observed in Table 2, it is not significant in Table 3. Again, this could be another indicator for the proximity to market’s weak predictive power compared to the other two measures examined in this paper.
The results for the interaction variables in Table 3 also exhibit similar features to those reported in Table 2. An exception to this is the coefficient of the interaction between PDS and proximity to the market, which becomes significant at the 1% level in Table 3. This again suggests that there is an issue of stability in using proximity to the market as a measure of market access compared to the alternative market access measures examined.
In addition to exploring the relationship between household dietary diversity and some selected control variables, we found a few other socioeconomic variables that influence household dietary diversity. Examining marital status, the finding shows that a single household head positively influences HDDS. However, caution should be taken in handling this result, because the number of single household heads in our survey was minimal (See Table 1). The other influencing factor is the geographic locations of the households, specifically the sub-district in which they are located. Damaino sub-district residents on average consume 0.621 lower food groups, (significant at 1%), than those in Geter Haiki-Mesihal, the reference category. Similarly, residents of Degaabur consume 0.819 food groups with lower household dietary diversity than households in Geter Haiki-Mesihal, significant at 1%. The education level of household heads positively affects household dietary diversity, at the 5% significance level. The other influencer of HDDS in our study areas was the size of household irrigable farmland. An additional one Tsimad (a quarter of a hectare) of irrigable land ownership leads to an average increase in HDDS by 0.320 – 0.595 food groups, significant at 1% -5% levels.
3.3.4. Robustness checks
The robustness of our findings is checked by applying different model fit tests. The first is the Likelihood Ratio (LR) Test, used to compare the fit of the Poisson model against alternative models such as the negative binomial regression to check for overdispersion. The result (LR test alpha = 0: chibar2(01) = 0.000; Prob > = chibar2 = 1.000) indicates that there is no evidence of overdispersion, and the Poisson model used is a good fit. In addition, we run the Pearson goodness of fit to see if our data fits the Poisson regression model. The Pearson test result chi2 = 83.31748; Prob > chi2(382) = 1.000, shows that the fitness of the data to our model is good.
This research aims to empirically investigate the discrete effects of, as well as the joint interplay between, farm production diversity and market access on household dietary diversity. The findings of this research indicate that farm production diversity and market access, jointly and independently, have a positive and significant nonlinear influence on rural smallholder household dietary diversity. The right mix between farm production diversity and smallholder households’ frequency of food market visits increases rural household dietary diversity. Whereas some studies (e.g. Sibhatu & Qaim, 2017) suggest greater market access can act as a substitute for lower farm production diversity in promoting household dietary diversity, this study has a greater emphasis on the positive and complementary contribution of both.
One policy implication is that farm production diversity that promotes farmers’ frequent market visits, such as through the production of short gestation high value vegetables, like tomatoes, kale, spinach, lettuce and green chillies, and poultry farming can enhance household dietary diversity. This approach can support both commercialization with frequent market attendance and self-consumption of diverse foods at the household level.
Promoting an expansion of small-scale irrigation would support increased production of these high value vegetable crops.
The other major policy implication arising from the study is to promote improved access to market through the development of rural road and related transport infrastructures. These measures implemented in tandem are feasible and promising approaches to support the joint positive impact on household dietary diversity.
FUNDING DETAILS
The research was supported by Teagasc- Agriculture and Food Development Authority, Ireland, as part of the Walsh Scholarship Program under grant number 2018013.
COMPETING INTERESTS
The authors have no relevant financial or non-financial interests to disclose.
AUTHORS’ CONTRIBUTIONS STATEMENT
The principal investigator has been involved in the conception and design, supervising the data collection, cleaned the collected data, analysing and interpretation of the data, and writing the draft paper. The respected co-authors have been actively involved in monitoring the research process from conception to the write up of the draft, revising critically the intellectual contents, suggesting, and approval of the final version to be published. All the authors agree to be accountable for all aspects of the work.
DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from Teagasc- Agriculture and Food Development Authority, Ireland, but restrictions apply to the availability of these data, which were used under license for the current study and so are not publicly available. The data are, however, available from the authors upon reasonable request. However, the data are not made public due to privacy and ethical restrictions.
ETHICS APPROVAL
The study has got approval by the College of Business and Economics Research and Community Services Office at Mekelle University, Tigray, Ethiopia on August 20/2020 with a reference number CBE/4019/T-38.
CONSENT TO PARTICIPATE
An informed consent was translated into the local language and read to every participating individual. Only the voluntarily participating households have consented to the survey and have been informed that the confidentiality of their personal information will be protected.
CONSENT FOR PUBLICATION
Research participants were consented that the findings of the study will be used for academic purpose and could be presented in conferences and may be published.
ACKNOWLEDGEMENT
This project was financially supported by the Teagasc Walsh Scholarship programme and Irish Aid, we also acknowledge the support made by Mekelle University, Tsirae Wemberta Wereda, and Tabia Hayelom district administration offices in Tigray, Ethiopia.
- Admasu, A., Meresa Kiros, and, & Abdulkadir Memhur. (2011). Baseline survey report of Tigray Region on WASH.
- Argaw, T. L., Euan Phimister, & Deborah Roberts. (2021). From farm to kitchen: how gender affects production diversity and the dietary intake of farm households in Ethiopia. Journal of Agricultural Economics, 72(1), 268–292.
- Bellon, M. R., Bekele Hundie Kotu, Carlo Azzarri, & Francesco Caracciolo. (2020). To diversify or not to diversify, that is the question. Pursuing agricultural development for smallholder farmers in marginal areas of Ghana. World Development, 125, 104682.
- Bergau, S., Loos, T. K., & Sariyev, O. (2022). On- and Off-Farm Diversification and Travel Time to Markets: Linkages to Food Security in Rural Ethiopia. The European Journal of Development Research, 34(5), 2543–2560. https://doi.org/10.1057/s41287-021-00475-0
- Berti, P. R. (2015). Relationship between production diversity and dietary diversity depends on how number of foods is counted. Proceedings of the National Academy of Sciences, 112(42). https://doi.org/10.1073/pnas.1517006112
- Bezu, D. C. (2018). A review of factors affecting food security situation of Ethiopia: From the perspectives of FAD, economic and political economy theories. Int. J. Agric. Innov. Res , 6, 2319–2473.
- Bonuedi, I., Kornher, L., & Gerber, N. (2022). Agricultural seasonality, market access, and food security in Sierra Leone. Food Security, 14(2), 471–494.
- Chamberlin, J., & Jayne, T. S. (2013). Unpacking the Meaning of ‘Market Access’: Evidence from Rural Kenya. World Development, 41, 245–264. https://doi.org/10.1016/j.worlddev.2012.06.004
- Chegere, M. J., & Stage, J. (2020). Agricultural production diversity, dietary diversity and nutritional status: Panel data evidence from Tanzania. World Development, 129, 104856. https://doi.org/10.1016/j.worlddev.2019.104856
- Deribe, R. A. H. E. L. (2008). Institutional analysis of water management on communal irrigation systems in Ethiopia: the case of Atsbi Wemberta, Tigray Region and Ada’a Woreda, Oromiya Region [PhD dissertation]. Addis Ababa University.
- Disha, A., Kuntal K., S., Phuong H., N., Michael T., D., Marie T., R., Purnima, M., & Rahul, R. (2013). Household Food Insecurity Is Associated with Higher Child Undernutrition in Bangladesh, Ethiopia, and Vietnam, but the Effect Is Not Mediated by Child Dietary Diversity. The Journal of Nutrition, 143(12), 2015–2021. https://doi.org/10.3945/jn.113.175182
- Dorosh, P., & Minten B. (2020). Ethiopia’s agrifood system: Past trends, present challenges, and future scenarios. https://doi.org/10.2499/9780896296916
- Eckhardt, C. L. (2006). Micronutrient malnutrition, obesity, and chronic disease in countries undergoing the nutrition transition: potential links and program/policy implications (583-2016–39710; 213 ).
- FAO, F. (2007). Guidelines for measuring household and individual dietary diversity. Food and Agriculture Organization of the United Nations (FAO) the Food and Nutrition Technical Assistance (FANTA) Project.
- FAO, IFAD, UNICEF, WFP, & WHO. (2023). The State of Food Security and Nutrition in the World 2023. FAO; IFAD; UNICEF; WFP; WHO; https://doi.org/10.4060/cc3017en
- Gebremedhin Woldewahid, Berhanu Gebremedhin, Kahsay Berhe, & Dirk Hoekstra. (2011). Shifting towards market-oriented irrigated crops development as an approach to improve the income of farmers: Evidence from northern Ethiopia.
- Gessesse, T. A. (2016). Above- and belowground carbon stocks in semi-arid land-use systems under integrated watershed management in Gergera watershed, Ethiopia [Inaugural-Dissertation ]. Rheinischen Friedrich-Wilhelms-Universität .
- Hailu, A. G., & Amare, Z. Y. (2022). Impact of productive safety net program on food security of beneficiary households in western Ethiopia: A matching estimator approach. PLOS ONE, 17(1), e0260817. https://doi.org/10.1371/journal.pone.0260817
- Headey, D., Hirvonen, K., Hoddinott, J., & Stifel D. (2019). Rural food markets and child nutrition. American Journal of Agricultural Economics, 101(5), 1311–1327.
- Hoddinott, J., Headey, D., & Dereje, M. (2015). Cows, Missing Milk Markets, and Nutrition in Rural Ethiopia. The Journal of Development Studies, 51(8), 958–975. https://doi.org/10.1080/00220388.2015.1018903
- International Center for Research on Agroforestry (ICRAF). (2018). mpact Assessment of Enhancing Integrated Watershed Management with Climate Smart Agriculture at Gergera Watershed in Tigray, northern Ethiopia.
- Jones, A. D. (2017). Critical review of the emerging research evidence on agricultural biodiversity, diet diversity, and nutritional status in low- and middle-income countries. Nutrition Reviews, 75(10), 769–782. https://doi.org/10.1093/nutrit/nux040
- Jones, A. D., Shrinivas, A., & Bezner-Kerr, R. (2014). Farm production diversity is associated with greater household dietary diversity in Malawi: Findings from nationally representative data. Food Policy, 46, 1–12. https://doi.org/10.1016/j.foodpol.2014.02.001
- Kihiu, E. N., & Amuakwa-Mensah, F. (2021). Agricultural market access and dietary diversity in Kenya: Gender considerations towards improved household nutritional outcomes. Food Policy, 100, 102004. https://doi.org/10.1016/j.foodpol.2020.102004
- Kissoly, L. D., Karki, S. K., & Grote, U. (2020). Diversity in Farm Production and Household Diets: Comparing Evidence From Smallholders in Kenya and Tanzania. Frontiers in Sustainable Food Systems, 4. https://doi.org/10.3389/fsufs.2020.00077
- Koppmair, S., Kassie, M., & Qaim, M. (2017). Farm production, market access and dietary diversity in Malawi. Public Health Nutrition, 20(2), 325–335. https://doi.org/10.1017/S1368980016002135
- Kumar, N., Harris, J., & Rawat, R. (2015). If They Grow It, Will They Eat and Grow? Evidence from Zambia on Agricultural Diversity and Child Undernutrition. The Journal of Development Studies, 51(8), 1060–1077. https://doi.org/10.1080/00220388.2015.1018901
- Matita, M., Chirwa, E. W., Johnston, D., Mazalale, J., Smith, R., & Walls, H. (2021). Does household participation in food markets increase dietary diversity? Evidence from rural Malawi. Global Food Security, 28, 100486. https://doi.org/10.1016/j.gfs.2020.100486
- Meaza, G. B. (2010). Socio-economic analysis of market oriented beekeeping in Atsbi Wemberta District of Eastern Zone, Tigray Region [PhD dissertation]. Mekelle University.
- Morrissey, K., Reynolds, T., Tobin, D., & Isbell, C. (2024). Market engagement, crop diversity, dietary diversity, and food security: evidence from small-scale agricultural households in Uganda. Food Security, 16(1), 133–147. https://doi.org/10.1007/s12571-023-01411-2
- Muthini, D., Nzuma, J., & Nyikal, R. (2020a). Farm production diversity and its association with dietary diversity in Kenya. Food Security, 12(5), 1107–1120. https://doi.org/10.1007/s12571-020-01030-1
- Muthini, D., Nzuma, J., & Nyikal, R. (2020b). Farm production diversity and its association with dietary diversity in Kenya. Food Security, 12(5), 1107–1120. https://doi.org/10.1007/s12571-020-01030-1
- Nandi, R., Nedumaran, S., & Ravula, P. (2021). The interplay between food market access and farm household dietary diversity in low and middle income countries: A systematic review of literature. Global Food Security, 28, 100484. https://doi.org/10.1016/j.gfs.2020.100484
- Romeo, A., Meerman, J., Demeke, M., Scognamillo, A., & Asfaw, S. (2016). Linking farm diversification to household diet diversification: evidence from a sample of Kenyan ultra-poor farmers. Food Security, 8(6), 1069–1085. https://doi.org/10.1007/s12571-016-0617-3
- Sibhatu, K. T., Krishna, V. V., & Qaim, M. (2015). Production diversity and dietary diversity in smallholder farm households. Proceedings of the National Academy of Sciences, 112(34), 10657–10662. https://doi.org/10.1073/pnas.1510982112
- Sibhatu, K. T., & Qaim, M. (2017). Rural food security, subsistence agriculture, and seasonality. PLOS ONE, 12(10), e0186406. https://doi.org/10.1371/journal.pone.0186406
- Sibhatu, K. T., & Qaim, M. (2018). Farm production diversity and dietary quality: linkages and measurement issues. Food Security, 10(1), 47–59. https://doi.org/10.1007/s12571-017-0762-3
- Swindale, A., & Bilinsky, P. (2006). Development of a Universally Applicable Household Food Insecurity Measurement Tool: Process, Current Status, and Outstanding Issues. The Journal of Nutrition, 136(5), 1449S-1452S. https://doi.org/10.1093/jn/136.5.1449S
- Tesfaye, W., & Tirivayi, N. (2020). Crop diversity, household welfare and consumption smoothing under risk: Evidence from rural Uganda. World Development, 125, 104686. https://doi.org/10.1016/j.worlddev.2019.104686
- UNDP, A. (2012). Africa Human Development Report 2012 Towards a Food Secure Future.
- Usman, M. A., & Haile, M. G. (2022). Market access, household dietary diversity and food security: Evidence from Eastern Africa. Food Policy, 113, 102374. https://doi.org/10.1016/j.foodpol.2022.102374
- Woldewahid G., Gebremedhin B., Hoekstra D., Tegegne A, Berhe K., & Weldemariam D. (2012). Market‐oriented beekeeping development to improve smallholder income: Results of development experiences in Atsbi‐Womberta District, northern Ethiopia.
- Zanello, G., Shankar, B., & Poole, N. (2019). Buy or make? Agricultural production diversity, markets and dietary diversity in Afghanistan. Food Policy, 87, 101731. https://doi.org/10.1016/j.foodpol.2019.101731
The authors declare no competing interests.