Edited by: Joyce Ludovick Kinabo, Sokoine University of Agriculture, Tanzania
Reviewed by: Aida Turrini, Council for Agricultural and Economics Research (CREA), Italy; Youssef Aboussaleh, Ibn Tofail University, Morocco
This article was submitted to Nutrition and Sustainable Diets, a section of the journal Frontiers in Sustainable Food Systems
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
Farm production diversity is widely promoted as a strategy for enhancing smallholders' food and nutrition security. Nonetheless, empirical evidence from the rural smallholder context is still mixed. This study compares the nature, determinants and influence of farm production diversity on household dietary diversity in rural and peri-urban settings in Kenya and Tanzania. Descriptive and econometric analyses are employed using household-level survey data from four counties in Kenya (
Enhancing smallholder farm production diversity has recently gained increased attention owing to its potential to enhance rural households' food and nutrition security (Fanzo et al.,
Agricultural diversification is among widely advocated strategies to address the above challenges (Powell et al.,
Numerous recent studies, however, acknowledge that the relationship between farm production diversity and dietary diversity is complex and inherently confounded by various other factors such as market access (Jones et al.,
The present study contributes to this literature by comparatively assessing the nature, determinants and role of farm production diversity on household dietary diversity using the case of agricultural households in Kenya and Tanzania. A few studies look at the relationship between farm production diversity and dietary diversity at sub-national levels (e.g., Herforth,
Against this background, this comparative study intends (1) to examine the nature and determinants of farm production diversity and (2) to analyze the influence of farm production diversity on household dietary diversity using the diverse market and agro-ecological contexts of Kenya and Tanzania. The rest of the study is organized as follows. The next section describes the study areas and data, while section three elaborates on the conceptual framework and methodology used in this study. Results and discussion are presented in section four. Section 5 summarizes the main results and gives concluding remarks.
This study uses household-level survey data from Kenya and Tanzania collected in 2014. For Kenya, the data was collected from four counties namely Kisii, Kakamega, Kiambu and Nakuru (See
Map of study area-Kenya. Source: HORTINLEA survey, 2014.
In Tanzania, data was collected from smallholders in two districts, Kilosa and Chamwino (See
Map of study area—Tanzania. Source: Trans-Sec survey, 2014.
A summary of key characteristics of the study areas is provided in
Summary of main characteristics of study area.
Climate | Semi-humid | Semi-humid | Semi-humid | Semi-arid |
Agricultural potential | Relatively good | Relatively good | Relatively good | Relatively poor |
Access to major markets | Relatively good | Relatively poor | Relatively good | Relatively poor |
Major crops: Food crops Cash crops | Maize, potatoes, vegetables Tea, coffee, pyrethrum) | Maize, vegetables Tea, coffee, sugarcane | Maize, rice, peas Sesame, cotton | Sorghum, millet, groundnuts Sunflower, sesame |
Livestock | Dairy cattle, sheep | Dairy cattle | Little livestock keeping(poultry, goats) | Heavy integration of livestock(Cattle, goat, poultry) |
In assessing and comparing the role of farm production diversity on household dietary diversity in Kenya and Tanzania, we conceptualize key relationships as follows (
Conceptual framework (Authors' construction based on Scoones,
Several studies have proposed and used various measures of farm production diversity and dietary diversity. Starting with farm production diversity, different measures have evolved from previous studies that focused on assessing genetic diversity at the farm and on biodiversity (Hawksworth,
Regarding dietary diversity, we use two indicators. The first is the Household Dietary Diversity Score (HDDS). HDDS is a good proxy indicator for diet quality and is documented to correlate well with important nutrition outcomes such as anthropometric status (Swindale and Bilinsky,
Deriving from the conceptual framework, farm production diversity is influenced by various livelihood assets such as human, natural, social, physical and financial capital. We therefore assess the determinants of farm production diversity using relevant count data models: Poisson and Negative Binomial regression models a regression model specified as:
where
Household dietary diversity is assumed to be influenced by farm production among other factors. To specifically analyze this relationship for Kenya and Tanzania, we also specify a regression model in which household dietary diversity is determined by farm production diversity and other important control variables (see the list and description of variables used in the
where
Apart from farm production diversity, household dietary diversity can be influenced by household socio-economic characteristics such as age and gender of the household head which may determine households' dietary preferences and allocation of household resources toward food consumption (Jones et al.,
As noted, both specified relationships above in equations (1) and (2) are estimated with count data models i.e., Poisson and negative binomial regression models owing to the nature of our diversity indicators. We first carry out over-dispersion tests in our dependent variables to ascertain the need for employing a Poisson or negative binomial regression. For equi-dispersion, Poisson regression is used while the negative binomial regression is used in case of over-dispersed count data. Also, potential collinearity among explanatory variables is tested. As the present study rely on cross-section data, it must be pointed out that the results enable us to only assess potential associations between our variables of interest. Therefore, caution should be taken when interpreting the results as they may not necessarily imply causation.
Descriptive statistics in
Descriptive statistics of key household and farm characteristics in Kenya and Tanzania [Mean(Standard deviation)].
Age (years) | Age of the household head | 49.71(12.49) | 48.64(17.10) |
Gender (Male =1) | Gender of the household head | 0.80(0.39) | 0.78(0.40) |
Education (Formal =1) | Household head has formal education | 0.73(0.44) | 0.67(0.47) |
Labor (Worker equivalents) | Labor capacity | 4.11(1.92) | 3.02(1.47) |
Land (ha) | Total land | 0.82(0.80) | 1.71(1.76) |
Rainfall (mm) | Mean annual rainfall | 1408.4(339.06) | 473.23(78.69) |
Distance (km) | Distance to the nearest major markets | 2.46(2.48) | 6.06(4.71) |
Assets (Score) | Household asset holding | 64.87(87.19) | 64.01(190.27) |
Market information (Yes = 1) | Access to market information | 0.38(0.48) | 0.45(0.47) |
Off-farm employment (Yes = 1) | Access to off-farm employment | 0.31(0.46) | 0.33(0.47) |
Non-farm self –employment (Yes =1) | Access to nonfarm self-employment | 0.18(0.38) | 0.25(0.43) |
Credit access (Yes =1) | Access to credit | 0.18(0.39) | 0.09(0.29) |
Observations | 1,150 | 899 |
In terms of diversity, results from
Comparison of farm production and dietary diversity in Kenya and Tanzania study areas [Scores: Mean (Standard deviation)].
Production diversity | 5.27 (1.38) | 4.34*** (1.52) | 4.96 | 3.01 (1.35) | 3.81*** (1.33) | 3.41 |
Dietary diversity | ||||||
HDDS | 6.28 (1.45) | 6.81 *** (1.30) | 6.46 | 5.29 (1.46) | 4.20*** (1.39) | 4.74 (1.52) |
FVS | 15.66 (4.08) | 18.64*** (5.27) | 16.66 | 10.95 (3.38) | 9.03 (3.82) | 9.99 (3.73) |
Observations | 766 | 384 | 1,150 | 450 | 448 | 899 |
Regression results of determinants of production diversity [coefficients (standard errors)].
Age (years) | 0.001 | 0.003** |
(0.002) | (0.001) | |
Gender (Male =1) | −0.027 | 0.069** |
(0.021) | (0.034) | |
Education (Formal =1) | 0.003 | 0.036 |
(0.021) | (0.029) | |
Labor (Worker equivalents) | 0.015*** | 0.031*** |
(0.003) | (0.008) | |
Land (ha) | 0.071*** | 0.029*** |
(0.010) | (0.006) | |
Rainfall (mm) | 0.001*** | −0.002** |
(0.000) | (0.000) | |
Distance (km) | 0.003 | 0.011** |
(0.002) | (0.005) | |
Assets (Score) | 0.001 | 0.001 |
(0.000) | (0.000) | |
Market information (Yes =1) | 0.041*** | 0.035 |
(0.01) | (0.027) | |
Off-farm employment (Yes =1) | 0.039** | 0.003 |
(0.018) | (0.028) | |
Nonfarm self–employment (Yes=1) | 0.019 | 0.075*** |
(0.020) | (0.027) | |
Credit access (Yes =1) | 0.036** | 0.112*** |
(0.019) | (0.029) | |
Risk attitude (Scale: 1–10) | −0.006* | 0.005 |
(0.004) | (0.005) | |
Shocks (Yes =1) | 0.049** | −0.058** |
(0.019) | (0.029) | |
Regional Dummy | −0.038** | −0.022** |
(0.024) | (0.015) | |
Constant | 1.122*** | 1.139*** |
(0.062) | (0.201) | |
Observations | 1,150 | 899 |
Wald chi2 | 203.20 | 226.03 |
Probability>chi2 | 0.000 | 0.00 |
Pseudo R2 | 0.04 | 0.031 |
In both countries, labor, land and credit access have a positive and significant contribution to farm production diversity. These constitute important household endowments which are critical in influencing the number of crops produced and livestock species kept by a household (Benin et al.,
As aforementioned, country-specific differences exist in how various factors influence farm production diversity. In Kenya, rainfall has a positive and significant effect on farm production diversity. The reason for this may be that, given the existing agro-ecological characteristics, availability of rainfall is likely to increase diversity maintained by smallholders, especially in terms of different crop species (Di Falco et al.,
Distance to the nearest major markets is significantly associated with increased farm production diversity only in Tanzania. This implies that smallholders in distant and less accessible areas tend to maintain higher levels of diversity in their farm production so as to circumvent higher transaction costs involved in acquiring food from markets (Benin et al.,
In terms of household financial capital, off-farm employment and non-farm self-employment are positively and significantly associated with farm production diversity. While off-farm employment is significant only for Kenya, non-farm self-employment is significant for Tanzania. Both are important sources of income to smallholders and they enable financing of various farm production operations such as inputs purchases. In Kenya, off-farm employment mostly takes the form of construction work or wholesale/retail trade (Khatri Karki,
With regards to other controls, results show that risk attitude could play a vital role in influencing farm production diversity in Kenya. Specifically, preparedness of a household to take risk has a negative and significant influence on farm production diversity. The reason for this may be that, smallholders who are more willing to take risks have a more specialized farm production portfolio as they aim at increasing efficiency and farm incomes. On the contrary, risk-averse smallholders are likely to maintain a more diverse farm production portfolio so as to reduce production risks (Di Falco and Chavas,
Results from the analysis of the relationship between farm production diversity and dietary diversity are presented in
Regression results of determinants of food consumption diversity (HDDS and FVS) [coefficients (standard errors)].
Production diversity | 0.022*** | 0.035*** | 0.031*** | 0.040*** |
(0.004) | (0.006) | (0.007) | (0.009) | |
Age (years) | −0.002*** | −0.003*** | −0.002*** | −0.003*** |
(0.001) | (0.001) | (0.001) | (0.001) | |
Gender (Male = 1) | 0.051*** | 0.029 | 0.012 | 0.004 |
(0.016) | (0.020) | (0.026) | (0.030) | |
Education (Formal = 1) | −0.009 | −0.020 | 0.030 | 0.041 |
(0.015) | (0.018) | (0.023) | (0.027) | |
Labor (Worker equivalents) | 0.017*** | 0.023*** | −0.005 | 0.006 |
(0.004) | (0.005) | (0.007) | (0.009) | |
Land (ha) | 0.017** | 0.003 | 0.009 | 0.010 |
(0.007) | (0.009) | (0.005) | (0.007) | |
Distance (km) | −0.003 | −0.005 | −0.016*** | −0.017*** |
(0.003) | (0.004) | (0.004) | (0.005) | |
Assets (Score) | 0.000*** | 0.000*** | 0.000** | 0.000** |
(0.000) | (0.000) | (0.000) | (0.000) | |
Livestock (TLU) | −0.001 | 0.004 | −0.001 | −0.002 |
(0.002) | (0.003) | (0.001) | (0.003) | |
Market information (Yes=1) | −0.007 | 0.008 | 0.085*** | 0.105*** |
(0.013) | (0.016) | (0.022) | (0.025) | |
Food consumption expenditure (PPP$) | 0.003*** | 0.004*** | 0.001*** | 0.001*** |
(0.001) | (0.001) | (0.000) | (0.000) | |
Off-farm employment (Yes=1) | 0.002 | −0.001 | −0.046** | −0.049* |
(0.013) | (0.016) | (0.023) | (0.026) | |
Nonfarm self-employment (Yes=1) | 0.072*** | 0.062*** | 0.043* | 0.055** |
(0.014) | (0.018) | (0.022) | (0.027) | |
Credit access (Yes=1) | 0.028* | 0.046** | 0.042 | 0.049 |
(0.015) | (0.019) | (0.030) | (0.038) | |
Regional dummy | 0.099*** | 0.193*** | 0.089** | 0.055* |
(0.015) | (0.018) | (0.042) | (0.049) | |
Constant | 1.590*** | 2.429*** | 1.492*** | 2.212*** |
(0.044) | (0.061) | (0.076) | (0.083) | |
Ln(alpha) | −4.336*** | |||
(0.419) | ||||
Observations | 1150 | 1150 | 899 | 899 |
Wald chi2 | 215.32 | 307.34 | 350.74 | 202.02 |
Probability>chi2 | 0.000 | 0.000 | 0.00 | 0.00 |
Pseudo R2 | 0.01 | 0.06 | 0.032 | 0.041 |
The relationship between farm production diversity and dietary diversity is complex (Jones et al.,
Market related factors are also important determinants of dietary diversity. Distance to nearest major markets influences dietary diversity negatively for the case of Tanzania. This suggests that, with limited access to markets and other essential services, smallholders are not only constrained in terms of accessing a variety of food items from markets but also lack essential support infrastructure to improve their agricultural production. Dietary diversity is also positively related to access to market information for both countries, Kenya and Tanzania. Similarly, Sibhatu et al. (
Dietary diversity is also significantly influenced by household income. Our results show that food consumption expenditure and access to non-farm self-employment have a positive and significant effect on household dietary diversity for both Kenya and Tanzania. Access to remunerative non-farm self-employment income adds to household incomes and thus raises the households' purchasing power. With increased purchasing power, households may spend on more diverse food and hence improve their dietary diversity. Several studies note the positive role of increased household food consumption expenditure resulting from various income generating activities. For example, Jones et al. (
Location characteristics have also significant influence on household dietary diversity. Being located in peri-urban counties (for Kenya) and those in Kilosa for Tanzania is positively associated with increased dietary diversity. With regards to Kenya, this may reflect the fact that households in peri-urban areas have more opportunities in terms of market access thus being able to sell their produce and also purchase different food items. For Tanzania, Kilosa district has more agricultural potential given its semi-humid agro-ecology and also has better market access thus impacting household dietary diversity positively unlike in Chamwino district which is semi-arid with low market access.
Results on the analysis of the potential of farm production diversity on the seasonal household dietary diversity are presented in
Regression results of determinants of seasonal dietary diversity [coefficients (standard errors)].
Production diversity | 0.007*** | 0.004 | 0.016*** | 0.024*** | 0.024*** | 0.011* |
(0.003) | (0.004) | (0.003) | (0.008) | (0.007) | (0.007) | |
Age (years) | 0.000 | 0.000 | −0.000 | −0.001* | −0.002** | −0.001** |
(0.000) | (0.000) | (0.000) | (0.001) | (0.001) | (0.001) | |
Gender (Male =1) | 0.005 | 0.003 | 0.003 | −0.001 | 0.058** | −0.024 |
(0.011) | (0.013) | (0.009) | (0.027) | (0.029) | (0.023) | |
Education (Formal =1) | 0.009 | 0.017 | 0.002 | 0.048* | 0.012 | 0.029 |
(0.010) | (0.013) | (0.009) | (0.026) | (0.025) | (0.022) | |
Labor (Worker equivalents) | −0.005* | 0.000 | −0.002 | −0.007 | −0.010 | 0.006 |
(0.003) | (0.003) | (0.002) | (0.007) | (0.008) | (0.006) | |
Land (ha) | −0.004 | 0.007 | −0.002 | 0.012** | 0.017*** | 0.009* |
(0.005) | (0.005) | (0.004) | (0.005) | (0.006) | (0.005) | |
Distance (km) | −0.004* | −0.003 | 0.000 | −0.005 | −0.007 | −0.006 |
(0.002) | (0.002) | (0.001) | (0.005) | (0.005) | (0.004) | |
Assets (Score) | 0.000** | 0.000* | 0.000* | 0.000*** | 0.000*** | 0.000 |
(0.000) | (0.000) | (0.000) | (0.000) | (0.000) | (0.000) | |
Livestock (TLU) | 0.002* | 0.002 | 0.002* | 0.002* | 0.003* | 0.002 |
(0.001) | (0.001) | (0.001) | (0.001) | (0.002) | (0.001) | |
Market information (Yes =1) | 0.004 | −0.003 | 0.019*** | 0.104*** | 0.098*** | 0.069*** |
(0.008) | (0.010) | (0.007) | (0.023) | (0.024) | (0.020) | |
Food consumption expenditure | −0.000 | 0.000 | 0.000 | 0.001*** | 0.001*** | 0.001*** |
(PPP$) | (0.000) | (0.000) | (0.000) | (0.000) | (0.000) | (0.000) |
Off-farm employment (Yes =1) | 0.004 | 0.027*** | −0.010 | −0.033 | −0.027 | −0.014 |
(0.008) | (0.010) | (0.008) | (0.024) | (0.024) | (0.019) | |
Nonfarm self-employment (Yes =1) | −0.003 | 0.008 | 0.009 | 0.018 | 0.025 | 0.064*** |
(0.009) | (0.012) | (0.008) | (0.023) | (0.024) | (0.019) | |
Credit access (Yes =1) | 0.014* | −0.014 | 0.001 | −0.021 | 0.005 | −0.004 |
(0.008) | (0.012) | (0.008) | (0.032) | (0.034) | (0.030) | |
Regional dummy | 0.060*** | 0.094*** | 0.057*** | 0.210*** | 0.188*** | 0.081** |
(0.009) | (0.011) | (0.008) | (0.044) | (0.043) | (0.036) | |
Constant | 2.052*** | 1.999*** | 2.029*** | 1.480*** | 1.471*** | 1.697*** |
(0.027) | (0.031) | (0.025) | (0.075) | (0.077) | (0.065) | |
Observations | 1150 | 1150 | 1150 | 899 | 899 | 899 |
Wald chi2 | 108.41 | 151.78 | 102.86 | 291.21 | 304.99 | 138.23 |
Probability>chi2 | 0.000 | 0.000 | 0.000 | 0.00 | 0.00 | 0.00 |
Pseudo R2 | 0.00 | 0.00 | 0.00 | 0.035 | 0.036 | 0.014 |
The present study assessed and compared the nature and determinants of farm production diversity and its influence on household dietary diversity in Kenya and Tanzania.
Comparing the level of farm production diversity in the two countries, results show that smallholders in Kenya have a higher diversity compared to their counterparts in Tanzania. However, in Kenya, smallholders in peri-urban counties that are closer to major markets are far less diverse when compared to those in rural counties. Similarly, in Tanzania, farm production diversity is low in villages with better market access and a higher agricultural potential compared to those with lower market access. Overall, households' endowments in human, natural, physical, social and financial capitals are found to be important factors influencing the level of farm production diversity.
With regards to dietary diversity, overall, households in Kenya have significantly higher diversity in their diets when compared to Tanzania. Nevertheless, results demonstrate a significant and positive association between farm production diversity and the indicators of household dietary diversity for both countries. We also find evidence of a positive role of farm production diversity for seasonal dietary diversity. In addition, apart from farm production diversity, factors such as household productive assets, access to off-farm income opportunities and market access are equally important in enhancing household dietary diversity. In particular, market access seems to play a critical role in enhancing dietary diversity.
In light of the above findings, several implications can be drawn from this study. First, maintaining a higher diversity in farm production can be beneficial for household dietary diversity. This may be applicable to diverse rural and peri-urban contexts with varying market access and agricultural potentials. Second, market related factors are equally important. Proximity to markets offer additional benefits for households: they are able to increase their dietary diversity through increased incomes from agriculture and off-farm opportunities and enhanced access to a diversified portfolio of food items from markets. In terms of policy, therefore, interventions geared toward improving smallholder households' dietary diversity should address both production as well as market-related challenges. Specifically, focus should be on addressing production related challenges especially in rural contexts with less market access. In addition, improvement of market institutions and infrastructure is important for enhancing dietary diversity in diverse contexts such as rural and peri-urban settings.
The datasets generated for this study are available on request to the corresponding author.
The paper is based on a chapter in the PhD thesis of LK (Kissoly,
LK, SK, and UG contributed to the conception and design of the study. LK and SK organized the database and performed the statistical analyses. LK wrote a first draft of the manuscript. SK and UG reviewed and substantially contributed to subsequent drafts of the manuscript. All authors contributed to manuscript revision, read and approved the submitted version.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Summary of variables used in the study.
Age | Age of the household head in years |
Gender | Gender of the household head (Male =1, otherwise 0) |
Education | Household head has formal education (Formal =1, otherwise 0) |
Labor | Labor capacity (Worker equivalents): captures labor available at the household, were calculated by weighting household members; <9 years =0; 9–15 =0.7; 16–49 =1 and above 49 years =0.7. |
Land (ha) | Size of agricultural land owned in hectors |
Rainfall (mm) | Mean annual rainfall in millimeters |
Distance (km) | Distance to the nearest major markets |
Assets (Score) | Household asset holding |
Market information | Household has access to market information through various channels (neighbor, village leader, radio, phone, etc.) |
Off-farm employment | Household has a member/members engaged in off-farm employment |
Non-farm self–employment | Household has a member/members engaged in non-farm self-employment |
Credit access | Household has access to credit (capturing financial capital at the household level) |
Risk attitude | Self-stated risk taking behavior (scale of 0 to 10) |
Shocks | Households face agricultural shock (1=Yes) |
HDDS | Household dietary diversity score (calculated from 9 different food groups consumed by a household in the previous normal week). |
FVS | Number of different food items consumed by a household in a given reference period |
1The Simpson's Index and the modified Margalef species richness index would have been alternative indicators but these are able to suitably capture only crop diversity (Di Falco and Chavas,