Consumption of protein-rich food items: effect of cattle ownership and land-use consolidation

Abstract

This study aimed to evaluate the frequency of protein-rich food consumption among impoverished rural households in Rwanda. Data from the Rwanda Institute of Statistics, gathered nationally from a random sample of 9,709 households in 2018, was utilized for a comprehensive analysis of the food security and vulnerability survey. Given the dependent variable’s nature, a diverse econometric approach was employed to identify the factors influencing the weekly consumption of key protein-rich foods (milk, meat, and beans) in Rwandan families. An instrumental variable technique was applied to assess the impact of a unified land use policy on the consumption of protein-rich foods in Rwandan households, considering the lack of a direct relationship with welfare outcomes in theory. The results indicate that land consolidation significantly influences the consumption of meat and pulses. Additionally, cattle ownership has a notable impact on milk and pulse intake. In light of these findings, we recommend that the government and development partners enhance support for farmers, particularly by providing subsidized farm inputs and increasing the distribution of cattle to eligible low-income households.

PUBLIC INTEREST STATEMENT

This study assessed what factors influence how much protein people eat. Results pointed out that land being grouped together for farming (land consolidation) actually lowered meat and milk consumption, while surprisingly raising the intake of beans and peas (pulses). This suggests people might keep cows for more than just food, like for fertilizer. Owning cattle and having more money also meant eating more protein. Interestingly, households led by women ate more protein than those led by men, and older people ate more than younger people. Families with more people and less education ate less protein. Living in a city and having more income sources or better social networks also meant eating more protein. This research suggests that policies to help farmers, especially women and those in rural areas, could improve protein intake for everyone in Rwanda.

Keywords:

• Protein consumption
• cattle ownership
• land consolidation
• heterogeneous ordered regression
• Rwanda

REVIEWING EDITOR:

• M. Luisa Escudero-Gilete, Nutrition and Bromatology, Universidad de Sevilla, Sevilla, Spain

JEL CODES:

• I38
• O13
• Q15

SUBJECTS:

• Agriculture; Agriculture and Food; Nutrition; Conservation - Environment Studies; Food Additives & Ingredients; Rural Development

1. Introduction

Livestock is seen as a foundation of benefits for households in rural areas whose livelihood is dependent on farming. Cattle increases food security by providing high-value protein, such as meat and milk, which are frequently scarce in the meals of the underprivileged (Rawlins et al., 2014). On a farm, livestock can assist in crop-related actions as well as providing the soil with biological manure and nutrients, which are cheap and effective fertilizers (see Kato et al., 2011). Emerging evidence suggests that an increased protein intake may contribute to a decreased risk of protein-energy malnutrition (Naeem et al., 2023).

Because of Rwanda’s landscape, terracing, and susceptibility to climate change’s negative effects (Calzadilla et al., 2013), a major problem is soil erosion that can be mitigated by the potential of organic manure to retain water (Lal, 2004). Organic manure is essential for maintaining agro-ecological systems in regions with extensive terracing because it promotes soil water storage year-round and improves water retention (Liu et al., 2013).

The productivity of agriculture is also favourably correlated with animal ownership. For instance, Pender et al. (2004) discuss methods to improve land degradation and agricultural output in Uganda. They demonstrate that agricultural productivity is lower in households with fewer livestock units. Additionally, cattle can be used by households as a shield against the risk of income earning for consumption smoothing.

Malnutrition poses a severe threat to public health, particularly in underdeveloped nations. Childhood and adult development are significantly impacted by malnutrition’s negative health effects (Tathiah et al., 2013). According to some estimates, children’s malnutrition accounts for more than 50% of fatalities in underdeveloped nations (Kandala et al., 2001). Despite its great economic and social growth in recent decades, Rwanda is one of the SSA nations that still have a high prevalence of malnutrition (Logie et al., 2008). The high malnutrition rate, particularly among poor and rural households, is a robust symptom of ongoing poverty and food insecurity (World Bank Group, 2018). According to recent figures, 20.6% of Rwandan families lack access to food, with the western province having the highest level of food insecurity at 36% (WFP, 2022).

Given that welfare outcomes should be embedded in agricultural development initiatives, the implications of cattle ownership and land use consolidation¹ on household welfare still need explanations in Rwanda. The documentation on the connection between development policies, more specifically agricultural policies, and welfare (food security) outcomes is still very crucial.

This research aims at assessing consumption frequency of protein-rich food items among rural poor households in Rwanda. It aims specifically to (1) examine the effect of livestock ownership and (2) scan the role of land use consolidation on the consumption of protein-rich food items. Besides, this study aims to (3) identify other determinants of protein-rich food consumption in Rwanda.

2. Literature review

2.1. Land use consolidation, crop productivity and farmers’ welfare

For a century, consolidation of rural land has been a crucial and effective strategy for development around the globe. Today, it is a crucial component of rural sustainable development and governments have allotted a significant portion of their annual budgets to rural development schemes in an effort to maintain arable land, increase agricultural output, boost rural socio-economic growth, and contribute to the creation of rural landscapes (Luo & Timothy, 2017). Land consolidation is necessary to promote rural development and advance the proficiency of land use. Since the middle of the 1990s, LC has been used in China in an effort to expand the area that may be used for agriculture, lessen fragmentation, and boost capacity (Zhang et al., 2014). Especially in rural settings, land consolidation is a useful tool for enhancing agricultural management and living conditions (Fritzsche, 1981).

Rural social and economic degradation in the nations of Central and Eastern Europe has gotten worse during the past twenty years. The adoption of new policies regarding the principles of land ownership and management is a requirement for reversing this decrease. It is anticipated that the need to address unfavorable land fragmentation and promote appropriate land use in conjunction with effective environmental solutions will lead to the development of new, sustainable land management systems throughout Europe (Pašakarnis & Maliene, 2010). It is vital to note that land use consolidation has been used as one of the solutions to deal with the economic issues caused by disjointed land use and reshape rural infrastructure in numerous regions of of the globe (Muhinda & Dusengemungu, 2013; Asiama et al., 2017; Munnangi et al., 2020).

While considering environmental and ecological issues, these measures should prioritise enhancing farm production and agrarian employment, taxation policy, and land rights protection legislation (Pašakarnis & Maliene, 2010). Land ownership fragmentation is a problem that needs to be addressed (Pašakarnis & Maliene, 2010). It was emphasized that the main objective of promoting equitable and sustainable global development is to eradicate poverty. Rural land consolidation seeks to reduce poverty while achieving effective and sustainable land usage (Cheng et al., 2021).

Policymakers have long sought to upgrade and intensify agriculture in Africa for increased productivity and food security (Muyombano & Espling, 2020). Munnangi et al. (2020) claimed that the consolidation of land has been seen as an effective land management method in comparison to fragmented land use in terms of farm profitability, land use, crop yield, as well as high-standing food security.

The majority of people in Sub-Saharan Africa’s rural communities depend mostly on their job in agriculture for their livelihood. This degree of agricultural development has been attributed to small-scale subsistence farming’s characteristics, which rely on the productive powers of the farm household and the employment of fundamental production procedures, most frequently rudimentary tools. According to Matunhu (2011), many households are vulnerable as a result of this production system and face food insecurity.

The extension of land, the increase in agricultural output, the enhancement of living and working conditions in rural regions, the reduction of ecological risk, and the support of rural development have all benefited significantly from land consolidation. As a result, it promotes rural development and fights against poverty. Additionally, it helps to address the issue of a lack of land, technology, and funding confronted by the progress of underprivileged areas and revitalize the rural economy. It also encourages the capitalization of land resources, opens up new opportunities for farmers to rise their incomes, and helps to create employment opportunities (Zhou et al., 2018).

Numerous research works were conducted to assess the success of land consolidation uptake in different countries around the world. According to Jin et al. (2017), there is little evidence that land consolidation increases agricultural productivity. Additionally, they discovered that the effectiveness of land consolidation initiatives showed distinct regional variations and that the driving forces were not compatible with the design of the policy. Consolidation of land use has improved the land market tools and increased the number of private farms with larger average acreages (Sagaydak & Sagaydak, 2022).

According to Du et al. (2018)’s analysis of the usefulness of land use consolidation in different regions of China, areas with significant productivity gains primarily benefited from an increase in the agricultural zone, while the project’s yearly yield per hectare fell shortly after consolidation. In terms of transportation methods of agricultural produce, agricultural mechanization level, farmers in different regions of China (Changsha, Guiyang, and Hangzh) expressed satisfaction with land consolidation projects in terms of mechanisms of land transfers in the hamlet and among the members of their families, apparent implication of land consolidation, amount of social insurance assistance, and involvement of rural people in agricultural cooperatives (Luo & Timothy, 2017).

Zhang et al. (2020) examined the ability of various rural land use consolidation models to reduce poverty and boost farm income. They discovered that land consolidation encourages the integration of human-land-industry structures in areas with high levels of poverty, thereby lowering the livelihoods of farming households’ vulnerability and encouraging them to acquire resources for sustainable livelihoods, which in turn helps to reduce poverty and thereby raise income. Cheng et al. (2021) experimentally investigated the effect of rural land consolidation on multi-dimensional poverty, using the difference-in-differences method and survey data from low-income families. They discovered that the livelihoods of poor households had significantly and favourably improved as a result of land use consolidation.

2.2. Land use consolidation in Rwanda

The Crop Intensification Programme (CIP), which aimed to boost the farm productivity of selected crops with potentials of high productivity and encourage joint exploitation of vast land areas was at the centre of Rwanda’s agricultural transformation. This programme was anticipated to result in significant economies of scale. A few of the critiques directed at this plan include the sole-cropping of a small number of specific crops, which negatively affects food security, and the increase in rural socioeconomic differentiation (Muyombano & Espling, 2020). Land use consolidation and the national land planning policy were introduced in Rwanda in 2008 in order to scale up land area for farm and non-farm investment and to prevent any land subdivision that would lead to parcels of size less than 1 hectare (Del Prete et al., 2019). As part of crop intensification programme, the goals of land use consolidation are to eradicate poverty and guarantee the safety of the nation’s food supply (Bizoza, 2021).

Kathiresan (2012) conducted the initial assessment of land use consolidation in Rwanda. According to hir findings, the associated production of selected crops under CIP has led to substantial gains in food production, including a 5-fold rise in maize output, a 3-fold increase in wheat and cassava, a 2-fold increase in Irish potato and beans, and a 30% increase in rice. In 2013, Muhinda and Dusengemungu (2013) noted that the productivity in consolidated land areas has subtly encouraged farmers to use inputs and extension services. The country’s overall food security situation has greatly improved as a result of CIP implementation. This accomplishment has been greatly enabled by land use consolidation.

Ndabikunze (2015) performed a study on 245 respondents who are members of three agricultural cooperatives in Nyanza District, Southern Rwanda. Farmers today benefit from production that is focused on market, savings, loan availability, among other things, as an outcome of the land consolidation initiative. Findings also show that the government supports land use consolidation by subsidizing agricultural inputs like quality pesticides and seed in required quantities. Regarding the role of land use consolidation, respondents agreed that it has numerous advantages, including an increase in agricultural yield, the promotion of resourceful and sustainable use of land resources, the protection of smallholder land rights, the protection and promotion of women’ access to land as well as women’s enjoyment of the advantages associated with new consolidation policies, and an improvement in farmers’ food security.

In contrast to the law, which mandates that optional involvement would be carried out based on discussions and democratic principles, Ntihinyurwa and Masum (2017) demonstrated a significant gap between the principles and their execution and draw attention to the coercive and compulsory engagement of local farmers by local authorities. Chigbu et al. (2019) examined the influence of land consolidation on food security among Rwandan households. The findings point to a beneficial association between the consolidation of land use and improved food status. In accordance with its aims of boosting crop output, and improving food security and enhancing rural livelihoods, Ntihinyurwa and Masum (2017) assessed the contribution of crop growers’ engagement in execution of land use consolidation policy.

Nilsson (2019) looked into the way land use consolidation affected the yield of small-scale agriculture. The research reveals a positive connection between land use consolidation and agricultural output, only for landholdings bigger than one hectare, which is much larger than the typical farm dimension. Because of further increases in their use amongst the consolidated farms seem to be allied with significant benefits, it further underscores the role of irrigation and non-organic fertilizers.

With respect to the effect of land use consolidation on price stability, Nsabimana et al. (2021) asserted that the prices of maize and beans in concentrated and non-concentrated zones are cointegrated and convergent. This shows that farmers who cultivated one CIP crop intensively were able to afford the market and buy diverse food items, and that price discrepancies among different zones have gradually diminished, perhaps improving the welfare of the CIP concentrated farmers.

Del Prete et al. (2019) evaluated the causal relationship between outcomes for food safety and nutrition sustenance and land consolidation under crop intensification programme. They discovered that, while involvement in land consolidation actions had a positive impact on roots and tubers’ consumption, it had an adverse influence on meat, fish and fruit consumptions as well as the probable accessibility of B12 vitamin in adopters’ diets.

Utilizing qualitative data gathered in the Nyabihu district of Rwanda, Mutangiza (2019) investigated the impact of land consolidation policy on the food and nutrition security of women potato growers. Findings revealed that the main institutional factors influencing the adoption of land use consolidation are availability of inputs, training opportunities, and a stable market for crops. The results also showed that consolidation increased the region’s food supply through an increase in potato output, and that it improved the well-being of women farmers. As a result of the land consolidation, women farmers shifted from subsistence to business farming and were able to access money and engage in other off-farm activities. The high expense of potato cultivation, however, prevented household food and nutrition security from being attained.

Singirankabo et al. (2022) empirically conducted an analysis of the associations between crop productivity on smallholder farmers and the security of land tenure. Results prove that despite posing a danger to the security of land tenure, the increase in small farm harvests was caused by a number of factors, primarily those related to the ongoing crop intensification programme.

2.3. Importance of cattle in socioeconomic development

For agricultural households in developing nations, livestock is crucial since it can serve as a revenue producer and a cushion against income risk to smooth spending (Nilsson et al., 2019). The importance of small-scale dairy farming is rising due to its potential to significantly support sustainable household livelihoods through financial stability, food security, and nutritional stability (Chagunda et al., 2016). Cattle distribution is one strategy to advance the food security and nutrition sustenance of the deprived people because to the rise in milk demand and consumption over the previous 20 years in low- and middle-income nations (Dominguez-Salas et al., 2019).

Hayami and Ruttan (1970) created a paradigm for the function of livestock ownership as resources for developing nations’ agricultural output. They demonstrate how livestock and land are examples of capital accumulation that predominantly encapsulate inputs from the agriculture sector.

There are a number of reasons why livestock is seen as a significant factor in the socio-economic expansion of the nations. Livestock is essential to rural livelihoods and the economies of emerging countries (Herrero et al., 2013). The primary purposes of livestock were to provide income for saving or building wealth, to cover expected and unforeseen needs, and to produce regular and supplemental revenue. Another important effect was the improvement in social standing, which rose with cattle size. If they want to sustain farmers’ lives and develop environmentally friendly agricultural systems, researchers, governmental organizations, and stakeholders should focus more on the shifting economic conditions of cattle and specifically buffalo production (Lambertz et al., 2012).

They act as a vital resource and safety net for the impoverished, especially women and pastoralist tribes. In addition, they offer a large amount of food for the majority of people in both urban and rural areas (Herrero et al., 2013). By giving impoverished people access to high value proteins that are usually in short supply in their diets, livestock enhances food security (Rawlins et al., 2014). On a farm, livestock can help with crop-related tasks and offer manure and nutrients in soil, which are cheap and effective fertilizers (Kato et al., 2011). In areas with considerable terracing, Zhang et al. (2013) claim that the addition of organic manure is critical for maintaining the viability of agro-ecosystems since it increases soil water storage capacity and improves water retention. Soil erosion is a significant problem in Rwanda because of the country’s geography, terracing, and exposure to climate change (Lal, 2004). Organic manure, in particular, has an influence on reducing the danger of soil erosion (Calzadilla et al., 2013).

Livestock have a significant social and economic role in agriculture because of their capacity to provide a variety of valuable goods, such as nutrient-rich food and income supplements for families. They also support the nation’s sociocultural and economic security (Kumar et al., 2015). The quantity of cattle owned by each household is positively correlated with the availability and consumption of animal derived food items like meat, milk and eggs. The size of cattle is linked to household wealth and income, which is also associated to animal health care use (Thumbi et al., 2015). Animal-derived foods are a good source of high energy, proteins, vitamins, and minerals. The nutritional status and dietary diversity of households can therefore be greatly enhanced by consuming animal food items, which has an effect on income levels, household production, and eventually national development, especially for rural people (Kariuki et al., 2013).

Not only livestock is essential for food security, but also for environmental protection (Ehui et al., 1998; Sansoucy, 1995). Especially in low-income settings and borderline habitats that are unsuitable for farming, livestock production is an important factor for the sustainable levels of food security in many nations. Animal sources account for around one-third of the protein ingested by people worldwide (Godber & Wall, 2014). Environmental issues brought on by the disposal of antibiotic and hormone use, as well as animal waste can be lessened with the use of specialized, streamlined, and concentrated livestock farms (Sekaran et al., 2021).

2.4. Cattle distribution under social protection programme in Rwanda

In Rwanda, livestock is regarded as a crucial component in reducing poverty. In Rwandan culture, cows in particular hold a significant place since they represent wealth, sign of respect, and social standing (Kimenyi, 1978). In the same way, gifting and getting a cow carries significant value and meaning (Gravel, 1967). The cow is revered in Rwandan culture for a number of reasons, including its significance as a symbol of dowry, recognition for a job well done, and, most importantly, as a key component of the clientele system during the monarchy (Hahirwa & Karinganire, 2017).

Girinka, also referred as the ‘one-cow per poor family’ programme, is a national pro-poor growth initiative in Rwanda that was started in 2006. Through dairy cattle husbandry and integrated crop cultivation, the programme seeks to lower the rate of chronic child malnutrition in the nation, improve household food security, and serve as an alternative source of income (Kim et al., 2011; Mazimpaka et al., 2017).

The goals of the Girinka programme also address a number of current agricultural challenges in Rwanda. For instance, according to estimates from the Food and Agriculture Organization (FAO) of the United Nations, 37% of Rwanda’s land requires soil retention before agriculture, 40% of the country’s land is highly prone to erosion, and the quality of the soil – including both organic matter and soil fertility – is rapidly declining (Government of Rwanda, 2009). In response, the Girinka programme offers its participants the ability to improve the soil quality on their field by using manure (Kim et al., 2011). Due to the country’s extremely dense population and scarcity of farming land (resulting in the exploitation of borderline fields and brief fallow seasons), Rwandan smallholder crop producers, particularly Girinka cow receivers, are pushed to maximize manure utilization.

Besides the Government of Rwanda, different non-governmental organizations were involved in cattle distribution in Rwanda, namely Send a Cow and Heifer International (Argent et al., 2014; Nilsson et al., 2019). Due to supply-side limits on the capacities for government and NGOs, some recipients received guidance along with the cow transfer, while other households just received the cattle revolving (Nilsson et al., 2019). Argent et al. (2014) revealed that asset transfers together with training have long-term and large economic consequences on milk production, livestock milk yields, household incomes, and asset buildup even in a location like rural Rwanda where links amid farmers and produce markets are weak. Some local social agencies who collaborate closely with MINAGRI and provide milk cows report that they do not offer training to recipients because they feel Rwandans are naturally skilled at raising cattle (Argent et al., 2014). Other barriers that recipients of the Girinka scheme experience have been identified. Among them are, for example, the implications of a semi-permanent drought, a deficiency of veterinary services, and a lack of lands for growing fodder (Kayigema & Rugege, 2014; Hahirwa & Karinganire, 2017), a shortage of or insufficient training of cow recipients on the management of the acquired cows (Argent et al., 2014), as well as inadequate instruction in the usage and management of manure (Kim et al., 2011).

According Kayigema and Rugege (2014), Girinka programme significantly contributed to the shift from traditional to better breeds, which in turn served to reduce poverty by retailing milk and utilizing manure in agricultural output. As it turns out, exotic breeds of cows produce more milk than traditional breeds do (Argent et al., 2014), which helps both the fight against infant malnutrition and the aforementioned generating of money. Nevertheless, the Girinka method of breeding cows is challenging because the distributed cows are frequently exotic breeds that are less resistant to tropical diseases and ‘long dry seasons’ brought on by negative effects of climate change (Hahirwa & Karinganire, 2017).

It is important to mention that the Rwandan cattle landscape is characterized by the predominance of local dairy varieties, exhibiting limited genetic diversity (Chagunda et al., 2018). This presents potential vulnerabilities to disease and environmental pressures. D’Andre et al. (2017) identified the indigenous Ankole breed as the most prevalent, followed by its crosses with exotic breeds. Despite constituting only 28% of the total cattle population, improved dairy breeds, primarily Holstein-Friesian, Jersey, Brown Swiss (concentrated in the Northern Milkshed), and lately Fleckvieh, contribute a remarkable 82% of the national milk production (Ngamije, 2022; MINAGRI, 2013). Notably, dairy cattle distribution is widespread, with crossbred and pure Holstein-Friesian breeds dominating the north and northwest regions (Mutimura et al., 2015). While crossbred cows enjoy wider preference, larger farms tend to prioritize pure Holstein-Friesian breeds for their enhanced milk production potential.

Further in Rwandan society, livestock ownership, particularly of dairy cows, transcends its mere economic function as a contributor to poverty reduction. It is deeply intertwined with cultural symbolism and social significance. Dairy cows act as potent signifiers of wealth, prestige, and social status within the community. Ezeanya (2014) highlights the profound value and meaning associated with the gifting and receiving of a cow in Rwandan culture, reflecting the historical association of milk as a prestigious food consumed primarily by the elite. This cultural embeddedness necessitates acknowledging the multifaceted role of livestock beyond its purely economic contribution, thereby enriching our understanding of its broader societal impact in Rwanda.

3. Materials and methods

3.1. Sources of data

The present study employed a comprehensive analysis of food security and vulnerability survey data collected by the Rwanda Institute of Statistics (NSR) in 2018, with a random sample of 9,709 households. Considering the nature of the outcome, we used a heterogeneous econometric approach to identify the factors affecting the weekly consumption frequency of the main protein-rich food items (milk, meat and beans) for the Rwandan households. Given that there is theoretically no direct effect of a unified land use strategy on welfare outcomes, we used an instrumental variable method to assess its effect on protein-rich food consumption among Rwandan households. The methods of data analysis also accounted for the disparities between the treatment and the comparison groups.

In this line, in order to eliminate interaction terms and cross-population comparisons of effects, which can be quite deceptive, heterogeneous choice models—also known as location-scale models—were found to be superior to logit/probit models in terms of improving the outcomes of econometric estimates (see Williams, 2009) (Table 1).

Table 1. Descriptive statistics and definitions of the study variables.

Variable	Obs	Mean (SD)	Definition
meat	9709	0.462 (1.146)	Number of days per week meat is consumed by the household
milk	9709	1.135 (2.248)	Number of days per week mil is consumed by the household
pulse	9709	5.668 (2.248)	Number of days per week pulses are consumed by the household
consolidation	7121	0.211 (0.408)	Part of the land is under land consolidation programme (1 = yes, 0 = if otherwise)
location	9709	0.161 (0.368)	Urban rural status of the household (1 = urban, 0 = rural)
food_sec	9709	3.207 (0.789)	Food situation of the household as measured by food consumption group (1 = severely food insecure, 2 = moderately food insecure, 3 = marginally food secure, and 4 = food secure)
wealth	8357	2.835 (1.363)	Household wealth index (1 = poorest, 2 = poor, 3 = medium, 4 = wealthy, and 5 = wealthiest)
age	9709	47.627 (15.318)	The age of household’s head in years
cattle	9709	0.413 (0.492)	Household owns or manages cattle (1 = yes, 0 = otherwise)
land size	9709	1.815 (1.583)	Size of the owned land, categories from 0 to 7 (0 = no land owned, …., 7 = 5 ha and above)
sex	9709	0.267 (0.442)	The gender of household head (1 = female, 0 = male)
education	9709	2.429 (1.493)	Level of education of household head (1 = no education …, 8 = completed university)
hhsize	9709	4.785 (2.08)	Number of the persons living in the household
liveli source	9709	0.798 (0.401)	Agriculture is the main source of household’s livelihood (1 = yes, 0 = otherwise)
liveli_acti	9709	2.023 (0.759)	Household livelihood activities (1 = one activities, 2 = two activities, 3 = three activities or more)
Social_network	9709	0.277 (0.448)	The head of household is a member of a cooperative or an association (1 = yes, 0 = otherwise)

3.1.1. Study variables and empirical model

The dependent variables (meat, milk, and pulse) measure the frequency of their consumption per week. The days in a week a food item is consumed implies that a surveyed household did not consume it or consumed it at least on one meal per day. It means that, for meat equals one for example, a surveyed household consumed meat on one day over the reference week. This category comprises households that consumed meat on one meal, on two meals, and even on three meals. It implies that distances among the categories are not fixed; they might be changing from one level to another (see Long & Freese, 2014).

In light of the characteristics of the dependent variable (measured by the number of the days meat, milk, or pulses are consumed along the week – ordered categories, which makes them categorical variables), an ordered logistic regression model was specified as per Eq. (1)(1) $$y_i={\alpha }_0+{\alpha }_1{x}_{i1}+{\alpha }_2{x}_{i2}+\dots +{\alpha }_k{x}_{\mathit{\text{ik}}}+{\sigma \epsilon }_i$$(1) . (1) $$y_i={\alpha }_0+{\alpha }_1{x}_{i1}+{\alpha }_2{x}_{i2}+\dots +{\alpha }_k{x}_{\mathit{\text{ik}}}+{\sigma \epsilon }_i$$(1) where the x’s constitute a vector of the independent variables; the α’s are coefficients to be estimated so as to examine the influence of each independent variable x on y, is in fact a distorted or restricted form of a latent variable, y∗, and ε_i is a residual term that is commonly supposed to have either a logistic or normal (0, 1) distribution, and σ is a parameter that enables the variance to be regulated upwards or downwards.

We do not actually estimate the α’s because y is not observed. Instead, we estimate so-called β’s parameters. As noted by Allison (1999, citing Amemiya (1985)), the relationship between the α’s and the β’s can be observed as per the Eq. (2)(2) $${\beta }_k={\alpha }_k/\sigma k=1,\dots ,K$$(2) . (2) $${\beta }_k={\alpha }_k/\sigma k=1,\dots ,K$$(2)

We now come to a possible issue with the ordered logit/probit model. The ratio between the β’s and the α’s is the same for all situations when is the same in all circumstances – residuals are homoscedastic. However, the ratio also varies when σ varies between cases due to heteroskedasticity (Allison 1999). According to Hoetker (2004, 17), naive evaluations of coefficients between groups might reveal differences where none exist, conceal differences that do, and even show differences in the opposite direction of what actually exists when there are even relatively small differences in residual variation.

Cross-group comparisons of effects are nonsensical if the residual variances are different because the scaling of coefficients will differ between groups. This is because, regardless of the variables included in the model, coefficients are always scaled to ensure that the residual variance is constant.

An econometric approach that accounts for these issues is the heterogeneous choice model. With this model, σ varies between cases, which corrects for the problem of heteroskedasticity. To do this, the heterogeneous choice model simultaneously fits two equations: one for the factors influencing the decision or outcome and another for the factors influencing the residual variance. Due to its more efficient and simple methods compared to other recommended alternatives, a heterogeneous choice (location-scale) model is the most commonly utilized in practice (Williams, 2008; Williams, 2010). Accordingly, an empirical heteroscedastic ordered logistic regression model of this study was specified as per Eq. (3)(3) $$y_i^{*}={\displaystyle \sum }_k{x}_{\mathit{\text{ik}}}{\beta }_k+{\epsilon }_i$$(3) . (3) $$y_i^{*}={\displaystyle \sum }_k{x}_{\mathit{\text{ik}}}{\beta }_k+{\epsilon }_i$$(3)

The value of the underlying latent variable is provided by the location/choice equation. For the ith observation in the example above, x is a vector of k values. The determinants of the choice or outcome are the x’s, which are the explanatory variables. The βs demonstrate how the Xs impact the decision. Following Williams (2010), the variance equation can be written as per Eq. (4)(4) $${\sigma }_i=\hspace{0.17em}\mathrm{\text{exp}}\hspace{0.17em}(\sum _j{z}_{\mathit{\text{ij}}}{\gamma }_j)$$(4) . (4) $${\sigma }_i=\hspace{0.17em}\mathrm{\text{exp}}\hspace{0.17em}(\sum _j{z}_{\mathit{\text{ij}}}{\gamma }_j)$$(4)

The complete heterogeneous choice model (with logit link) for an ordered variable y with M categories coded 1 to M (Williams, 2010) can thus be stated as per Eq. (5)(5) $$\begin{array}{c}P(y_i>m)=\mathit{\text{inv log}}\hspace{0.17em}\mathit{\text{it}}\left\{\frac{{{\displaystyle \sum }}_k{x}_{\mathit{\text{ik}}}{\beta }_k-{\kappa }_m}{\hspace{0.17em}\mathit{\text{exp}}\hspace{0.17em}\left({{\displaystyle \sum }}_j{z}_{\mathit{\text{ij}}}{\gamma }_j\right)}\right\}\\ =\mathit{\text{inv log}}\hspace{0.17em}\mathit{\text{it}}\left(\frac{{{\displaystyle \sum }}_k{x}_{\mathit{\text{ik}}}{\beta }_k-{\kappa }_m}{{\sigma }_i}\right),\hspace{0.17em}\text{m}=1,\hspace{0.17em}2,\hspace{0.17em}\dots .,M-1,\end{array}$$(5) . (5) $$\begin{array}{c}P(y_i>m)=\mathit{\text{inv log}}\hspace{0.17em}\mathit{\text{it}}\left\{\frac{{{\displaystyle \sum }}_k{x}_{\mathit{\text{ik}}}{\beta }_k-{\kappa }_m}{\hspace{0.17em}\mathit{\text{exp}}\hspace{0.17em}\left({{\displaystyle \sum }}_j{z}_{\mathit{\text{ij}}}{\gamma }_j\right)}\right\}\\ =\mathit{\text{inv log}}\hspace{0.17em}\mathit{\text{it}}\left(\frac{{{\displaystyle \sum }}_k{x}_{\mathit{\text{ik}}}{\beta }_k-{\kappa }_m}{{\sigma }_i}\right),\hspace{0.17em}\text{m}=1,\hspace{0.17em}2,\hspace{0.17em}\dots .,M-1,\end{array}$$(5) where inv log it =inverse logit function of x = exp(x)/{1 + exp(x)}, K₀ = ∞, K_m = −∞, $\mathit{\text{exp}}\left({\displaystyle {\sum }_j{z}_{\mathit{\text{ij}}}{\gamma }_j}\right)=\mathit{\text{exp}}\hspace{0.17em}\left\{\mathit{\text{ln}}({\sigma }_i)\right\}={\sigma }_i$.

3.1.2. Method of estimation

We have specified a heteroscedastic ordered logistic model as per the Eq. (3)(3) $$y_i^{*}={\displaystyle \sum }_k{x}_{\mathit{\text{ik}}}{\beta }_k+{\epsilon }_i$$(3) to assess the effect of cattle ownership and land use consolidation on the consumption of protein-rich food items in Rwanda. Following Cameron and Trivedi (2010), we employed an instrumental variable estimation method, given that the cattle ownership and land use consolidation would have no direct influence of household welfare.

4. Results and discussion

4.1. Results from econometric estimations

Using weekly consumption of protein-rich food items (meat, milk, pulse) as dependent variables, are reported in Table 2. Ordered logistic regression estimates are included for comparison with heteroscedastic ordered logistic regression estimates. Considering the LR chi2, the heteroscedastic ordered logistic regression estimates are more reliable than simple ordered logistic regression estimates considering that LR chi2 for the former is greater than LR chi2 for the latter: that is, 1,223.67 against 1,258.70 for milk, and 2,580.80 against 2,581.69, respectively. Unlike meat, milk consumption is negatively influenced by land use consolidation. Besides, cattle ownership has highly negative and very highly positive and significant influence on the consumption of meat and milk, respectively. Concerning pulse consumption, we observed a positive and significant effect of both land consolidation and cattle ownership.

Table 2. Effect of cattle ownership and land consolidation on the consumption of food items with high-value proteins.

Variable	Ordered logistic regression estimations			Heteroscedastic ordered logistic regression estimations
	Model 1 (Meat)	Model 2 (Milk)	Model 3 (Pulse)	Model 1 (Meat)	Model 2 (Milk)	Model 3 (Pulse)
	Coefficients (Std. dev.)	Coefficients (Std. dev.)	Coefficients (Sd. Dev.)	Coefficients (Std. dev.)	Coefficients (Std. dev.)	Coefficients (Sd. Dev.)
consolidation	−0.050 (0.085)	−0.045 (0.077)	0.211*** (0.073)	0.008 (0.110)	−0.013 (0.113)	0.249*** (0.085)
cattle	−0.192*** (0.073)	0.792*** (0.067)	0.105* (0.058)	−0.207** (0.102)	0.528*** (0.092)	0.105 (0.064)
age	−0.010*** (0.003)	−0.002 (0.002)	.008*** (0.002)	−0.010*** (0.003)	−0.003 (0.003)	0.008*** (0.002)
sex	−0.096 (0.097)	0.058 (0.084)	−0.087 (0.070)	−0.100 (0.097)	0.069 (0.098)	−0.086 (0.070)
education	0.115*** (0.026)	0.172*** (0.025)	−0.067*** (0.026)	0.113*** (0.026)	0.196*** (0.029)	−0.068** (0.026)
hhsize	−0.036** (0.018)	0.020 (0.016)	−0.062*** (0.015)	−0.036** (0.018)	0.018 (0.019)	−0.063*** (0.015)
land_size	0.092*** (0.026)	0.07*** (0.024)	0.030 (0.024)	0.092*** (0.026)	0.086*** (0.028)	0.032 (0.024)
wealth	0.401*** (0.031)	0.271*** (0.027)	0.013 (0.026)	0.400*** (0.032)	0.323*** (0.034)	0.013 (0.026)
food_consgroup	0.746*** (0.059)	0.739*** (0.051)	1.839*** (0.045)	0.745*** (0.063)	0.871*** (0.064)	1.852*** (0.056)
liveli_acti	0.253*** (0.049)	0.025 (0.043)	0.136*** (0.038)	0.253*** (0.049)	0.023 (0.050)	0.136*** (0.039)
liveli_source	−0.053 (0.084)	−0.179** (0.075)	−0.128** (0.065)	−0.054 (0.084)	−0.204** (0.088)	−0.129** (0.065)
social_network	0.048 (0.072)	0.029 (0.065)	0.118* (0.063)	0.049 (0.071)	0.021 (0.076)	0.121* (0.063)
location	0.367*** (0.122)	1.038*** (0.108)	0.047 (0.130)	0.366*** (0.122)	1.118*** (0.123)	0.047 (0.130)
lnsigma consolidation	–	–	–	−0.044 (0.055)	−0.045 (0.056)	0.041 (0.044)
lnsigma cattle	–	–	–	0.011 (0.048)	0.297 (0.050)	−0.001 (0.034)
/cut1	5.518 (0.274)	5.438 (0.242)	1.250 (0.204)	5.510*** (0.300)	6.062*** (0.308)	1.265*** (0.210)
/cut2	6.277 (0.277)	5.696 (0.243)	2.005 (0.195)	6.265*** (0.312)	6.371*** (0.313)	2.025*** (0.204)
/cut3	7.240 (0.283)	6.095 (0.245)	2.934 (0.192)	7.223*** (0.330)	6.851*** (0.322)	2.962*** (0.206)
/cut4	8.021 (0.290)	6.510 (0.247)	4.018 (0.194)	8.001*** (0.348)	7.355*** (0.332)	4.053*** (0.214)
/cut5	8.553 (0.299)	6.678 (0.248)	4.796 (0.197)	8.530*** (0.364)	7.560*** (0.337)	4.837*** (0.223)
/cut6	8.977 0.309)	6.787 (0.248)	5.275 (0.200)	8.952*** (0.379)	7.695*** (0.340)	5.319*** (0.230)
/cut7	9.112 (0.313)	6.827 (0.249)	5.621 (0.201)	9.086*** (0.384)	7.743*** (0.341)	5.668*** (0.234)
Observations	6486	6486	6486	6486	6486	6486
LR chi2	1223.67	1223.67	2580.80	907.37	1258.70	2581.69
Prob > chi2	0.000	0.000	0.000	0.000	0.000	0.000

*** p < 0.001, ** p < 0.05, * p < 0.10.

Beside cattle ownership, meat consumption is very highly influenced negatively by the age of the household head and the household size. It is very highly influenced positively by the level of education of the household head, the owned land size, the wealth category, the food security status, the number of household livelihood activities and the location of the household in urban areas. Another factor that impacts on the consumption of meat is the family size with highly significant and negative influence. For milk consumption, other determining factors than cattle ownership are the household head’s education level (very highly positive effect), the owned land size (very highly positive effect), the wealth (very highly positive effect), the food security status (very highly positive effect), the main livelihood sources (very highly negative effect), and the location (very highly positive effect).

As for pulse consumption, additional significant determinants are the household head’s age (very highly positive effect), the household head’s education level (highly negative effect), family size (very highly negative effect), land size (positive effect), wealth status (positive effect), food security status (very highly positive effect), livelihood activities (very highly positive effect), and social network (significantly positive effect). This indicates that socio-economic standing of the household, the wealth and ownership of assets, location and social capital are the primary factors determining the consumption of protein-rich food items.

4.2. Discussion of the findings

Meat and milk consumptions are negatively influenced by land use consolidation, with a significantly negative outcome of livestock possession. This implies that the adoption of land consolidation adoption and cattle ownership both reduce the likelihood that meat and milk will be consumed. It might also imply that the vast majority of Rwandan households raise cattle for the purpose of providing manure to boost crop productivity. Further, given that the majority of farmers are poor, cattle serve to supplement their income and help them smooth the consumption of the stapple food items, which are primarily tubers, grains, and pulses. We observed a positive effect of land consolidation on pulse consumption as well as a positive and significant effect of cow ownership. These results support Maniriho et al. (2021) who proved positive influence of livestock on protein consumption, as well as Monirul Alam et al. (2018) who claimed that adopting cattle has a very highly significant positive effect on household food security. In contrast, they oppose Herrera et al. (2021) who stated that animal ownership did not significantly affect family food insecurity. However, this finding opposes that of Nadi et al. (2021) who showed that a rise in the distribution of livestock assets negatively affected the household’s food security. Concerning the positive effect of land consolidation on protein consumption, our findings agree with Tran and Van Vu (2020), who stated that the more scattered the family’s farmland is, the higher the likelihood for the family to experience food insecurity.

Results indicate that the household head’s age is another substantial determining factor of protein-rich food consumption with a very highly positive effect. This implies that the higher the household head’s age, the more protein-rich foods items are consumed. This finding is correlative to that of Maniriho et al. (2021) who noted that the family head’s age positively influences a household’s likelihood of consuming protein, as well as Oluwatayo (2008) who urged that the family head’s age positively affects food security. Our results also support the claim made by van Phuong et al. (2014), Akinsulu et al. (2019), and and Matavel et al. (2022) that the consumption of meat (pork and poultry) declines with increasing household head age. However, this closure contradicts Titus and Adetokunbo (2007) report that the likelihood of food insecurity rises as the household heads’ ages.

Results show a very highly negative effect of the household head’s sex on the protein-rich food consumption among the households. It implies that the female-head families consume more protein-rich food items than those headed by men. This finding corroborates the assertion Titus and Adetokunbo (2007) and Maniriho et al. (2021) who highlighted that female-headed families are more likely to experience food insecurity compared to families led by male heads (Titus & Adetokunbo, 2007; Oluwatayo, 2008; Maniriho et al., 2021; Acheampong et al., 2022). It also supports van Phuong et al. (2014) who underscore that female-headed households consume less meat (pork and poultry) than male-headed ones do, as well as Akinsulu et al. (2019) who reported that men spend more money on meat consumption in comparison to female consumers. Further, this finding approves Matavel et al. (2022) notice that there exists a statistically significant connection of food security metrics with the household head’s sex.

Our results disclose a very highly negative effect of family size on meat, milk and pulse consumptions among households. It implores that food insecurity is more likely to occur in large households than in the small ones. This finding is compliant with that of Oduniyi and Tekana (2020) who announced that household size has a significantly adverse impact on food security, denoting that the likelihood of achieving food security is more likely to decline as household size. It also supports Titus and Adetokunbo (2007) who documented that the rate of food insecurity rises as household size grows, as well as Herrera et al. (2021) who set forth that families with more children are more likely to struggle with hunger, especially those with smaller landholdings. However, this finding is discordant with that of Sah et al. (2017) and Maniriho et al. (2021) who denoted that, besides other household socioeconomic characteristics, the household likelihood of consuming protein-rich food items is positively correlated with family size.

We found education as another primary determinant of meat, milk, and pulse consumption. In accordance with this finding, the people with more education are better equipped to sustain the food security of their households. This is likely due to improved nutrition knowledge and better access to food thanks to increased income levels. Thus, the combined impacts of education on household food security are anticipated to include improved nutrition knowledge and simpler access to food as a result of increased income (see De Cock et al., 2013). Accordingly, education significantly influences food security (see Ogunniyi et al., 2021; Maniriho et al., 2021).

A significantly important impact of owned land size and the consumption of protein-rich food items implies that farm output increases with farm size, and, consequently, the surplus generates income used by farmers to buy different food items not produced on their farms. This finding is supporting Herrera et al. (2021) who proclaimed that the incidence of poverty and hunger declines with larger farms, and thus disclosed that even huge households with larger arable lands experience lower levels of food insecurity. It comes also to endorse Omotesho et al. (2008) observation that the farm size is one of the key predictors of a household’s food sustenance, and that of Aidoo et al. (2013) who found that the household food situation significantly benefited from the farm’s size. This finding also leads to authenticate Sah et al. (2017) confirmation that rural household pulse consumption is seen to be significantly and significantly impacted by exploited land holding, as well as Matavel et al. (2022) who maintained that farm size and food security indices are statistically significantly correlated.

For the wealth category, our results reveal that it has a very highly positive effect on the consumption of meat, milk, and pulses, which implies that indicates that a better wealth category corresponds to household improved status of food sustenance. This finding matches Hjelm et al. (2016) declaration that food insecurity is primarily caused by poverty since it prevents poor households from having access to the minimum food quality required to maintain a healthy and active lifestyle. The finding is also congruent with Guo’s (2011) conviction that household’s assets considerably affect food and nutrition situation given that they avail necessary resources to regulate food consumption. It further supports Chen et al. (2023) who divulged that having more assets decreases the likelihood of suffering from poverty and hunger that can get worse or improve over time.

We remarked a strong positive connection of the household’s food consumption category with the consumption of protein-rich food items. It entails that the higher consumption score implies that a household is more food secure. This finding is congruent with Omotesho et al. (2008) finding that one of the key factors affecting a household’s degree of food security is the amount of money spent on food. Our finding is also in agreement with Sah et al. (2017) who revealed a positive correlated between food intake and household’s food status.

Our findings indicate that the diversity of economic activities has a very highly positive implications for the weekly consumption of meat, milk, and pulses in households. This suggests that household intake of food items with high-value proteins increases as the number of sources of income increases, which may be attributable to the variety of household income. This result supports Mensah’s (2014) assertion that the number of economic activities had a substantial effect on household’s food consumption. In a similar vein, it is consistent with Iram and Butt (2004) findings that income diversity boosts food consumption in diversified households while underlining the statistically significant impact that income diversification had on households’ calorie intake. Our findings also confirm Maniriho et al. (2021) report that protein consumption depends significantly on the diversity of the livelihood sources. Furthermore, our results support the claim made by Zeleke et al. (2017) that the diversification in household income had a significant impact on the quantity of calories consumed.

Our findings also prove the social network to affect positively the consumption of food items rich in proteins (meat, milk, and pulses). Participation in any organization has a substantial impact on food security (Kassie et al., 2014; Ogunniyi et al., 2021). According to Kehinde and Kehinde (2020) study, joining a cooperative significantly and favourably affects the food security of rural households. Nosratabadi et al. (2020) demonstrated the role of social capital in improving food security, and thus explained two mechanisms through which knowledge sharing and sharing food products among diverse households.

Results of this study substantiate that the household’s location has a very highly positive effect on the consumption of protein-rich food items. This implies that urban households consume more protein-rich food items on average than rural households do. This finding is congruent with Iram and Butt’s (2014) claim that urban households consume more calories per capita than rural households. It is also in line with Zhou et al.’s (2017) assertion that rural households are among the most vulnerable groups to food security. Further, according to Monirul Alam et al. (2018), rural households are more susceptible to poverty and hunger as an outcome of a limited access to numerous essential services and products, which can lead to a depressing cycle of poverty. It is in this way, Betru and Kawashima (2009) reported urbanization as one of the primary factors that positively and significantly influence meat consumption among households. Further, due to geographic restrictions that limit exchanges and encourage consumption of locally produced foods, Leterme and Muũoz (2007) clarified that the rural population consumes more pulses than the urban population.

Furthermore, this study’ findings are aligned with some sustainable development goals (SDGs), namely SDG2 (‘Zero Hunger’), SDG 1 (‘No Poverty’), SDG 12 (‘Responsible Consumption and Production’), SDG 5 (‘Gender Equality’), and SDG 8 (‘Decent Work and Economic Growth’), respectively. While consolidation enhances pulse consumption, it reduces meat and milk intake, highlighting the need for balanced strategies in food systems. This aligns with SDG 2 (‘Zero Hunger’), emphasizing the need for sustainable food systems that balance productivity with access to diverse dietary sources.

Household characteristics influence protein consumption, aligning with SDG 2 target 2.1 (‘End hunger and all forms of malnutrition’). In this line, age and education positively impact protein intake, suggesting the importance of knowledge and access. Similarly, the negative impact of family size supports SDG 2’s focus on vulnerable populations, as larger families face increased food insecurity risks. In addition, the positive correlation between land size, wealth category, and food consumption aligns with SDG 2’s target 2.3 (‘Double the agricultural productivity and income of small-scale food producers’). Supporting small-scale farmers through land access and income diversification programs can address rural poverty and food insecurity (SDG 1 ‘No Poverty’) while promoting sustainable agricultural practices (SDG 12 ‘Responsible Consumption and Production’). In the same vein, livelihood diversification and social networks’ positive impact align with SDGs 1, 2, and 5, suggesting that strengthening rural livelihoods and social capital can be crucial for food security and well-being.

5. Conclusion and policy recommendations

This investigation delved into the multifaceted determinants of protein consumption in Rwandan households, focusing on the interplay between land-use consolidation, cattle ownership, and various socioeconomic factors. Our findings unveiled a nuanced picture, revealing both positive and negative influences on the intake of meat, milk, and pulses, which comprise vital protein sources. The consolidation of land holdings exhibited a complex relationship with protein consumption. While it positively impacted pulse intake, it surprisingly coincided with a decrease in meat and milk consumption. This suggests potential alternative motives for cattle ownership beyond consumption, such as manure production for agricultural enhancement. Additionally, poverty emerged as a crucial motivator for raising cattle, serving as a source of income and facilitating the consumption of staple tubers, grains, and pulses.

Results indicated that family composition and demographics played a significant role. Older household heads fostered higher protein consumption, implying a potential generational shift in dietary priorities. Interestingly, households led by women displayed a statistically significant preference for protein-rich foods compared to those headed by men. Family size demonstrated a negative correlation with protein intake, highlighting the challenges faced by larger households in maintaining food security. Education, conversely, exerted a positive influence, likely due to improved nutritional knowledge and income-driven access to diverse food sources. Farm size and wealth category demonstrated positive associations with protein consumption. Larger farms facilitated surplus production, enabling the purchase of protein-rich foods not cultivated on the land. Similarly, a higher wealth category correlated with improved food security and increased protein intake. Notably, food consumption category itself emerged as a strong predictor of protein consumption, signifying the cumulative effect of varied dietary resources.

Represented by the number of income sources, livelihood diversification exhibited a positive relationship with protein intake. This suggests that multiple income streams enhance household access to a wider range of food items, including protein sources. Social networks also positively influenced protein consumption, highlighting the potential role of community support in securing dietary needs. Our analysis further revealed a significant urban-rural disparity in protein consumption, with urban households generally consuming more protein-rich foods than their rural counterparts. This underscores the need for targeted interventions to address potential food insecurity and dietary deficiencies in rural areas.

Based on these findings, we recommend policy interventions emphasizing technical and financial support for farmers, including subsidized farm supplies and targeted cattle distribution programs for vulnerable households. Furthermore, promoting the sustainable management of agricultural resources through optimized fertilizer and pesticide use, coupled with a review of agricultural extension services, should be prioritized to encourage food security and healthy dietary practices across Rwanda.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

The author acknowledges the financial support of Erasmus+ through Christian Albrechts University (CAU) that facilitated his research visit at Kiel Institute of the World Economy (IfW) in April–May 2023, which resulted in the manuscript of this paper.

Notes on contributors

Aristide Maniriho

Aristide Maniriho is a lecturer in the Department of Economics, School of Economics at University of Rwanda. He holds a PhD in Agricultural Economics and Rural Development from the University of Liege – Gembloux Agro-Bio Tech (Belgium) obtained in December 2021. Maniriho’s primary research thrust focuses on agricultural and development economics. This extends to a keen interest in environmental and natural resource economics, as well as food economics. Maniriho’s secondary research interests include university pedagogy of economics, and international economics, which reflects a nuanced understanding of the interconnectedness of various factors influencing agricultural and rural development. Crucially, Maniriho’s research aligns with broader themes of agricultural performance, sustainable development, poverty reduction, and rural development. His findings actively inform policy interventions and development programs, aiming to enhance food security and nutritional well-being in rural communities.

Notes

1 Livestock (both small and big animals) have been distributed among Rwandans through different schemes such as one cow per poor household programme (known as Girinka). Similarly, in 2006, the Government of Rwanda initiated a crop intensification programme (land use consolidation being one of its components) that aimed at boosting crop productivity of six most important crops for food security: bean, potato, rice, cassava, wheat and maize.