1 Introduction

In sites where agriculture production is common, soil erosion and nutrient depletion are typical, and without appropriate measures to maintain soil fertility, the land often becomes unproductive [1]. To counter the negative effect of the soil and nutrient loss, farmers tend to invest in agricultural and sustainable land management (SLM) practices that have the potential of improving land productivity [2]. Among the agricultural and SLM that farmers are investing on, soil carbon enhancing practices (SCEPs)Footnote 1 (such as compost manure, conservation tillage, contour planting, crop residues, crop rotation, drainage ditch, ‘fanya juu’ terraces, farmyard manure, grass strips etc.) are some of the most common ones [3].

Nearly 80% of the carbon found on the land ecosystem is stored in soils, thus the implementation of SCEPs proficiently helps in the sequestration of atmospheric carbonFootnote 2 in the form of soil organic carbon (SOC). Thus, SCEPs play a significant role in the mitigation of the effects of climate change [4, 5]. Further, SOC is a major element of soil organic matter (SOM) which improves soil structureFootnote 3 through nutrient release, thereby ensuring the sustainability of soil functions [6]. These, in turn, promote crops growth and quick regeneration of forages and hence livestock production. The adoption of SCEPs, therefore, has the potential of improving farm productivity, farmers’ income and food security [7]. However, despite the importance of SCEPs, their adoption in the eastern Africa region is still low [8, 9].

The literature on adoption of agricultural management and SLM practices that enhance sequestration of soil organic carbon or that mitigate the loss of soil carbon—hereafter referred to as SCEPs—is extensive [10, 11] and shows that the decision to adopt these practices is influenced by the socio-economic characteristics of the households, biophysical, plot or farm and institutional characteristics in diverse ways as influenced by the context and the environmental conditions [12,13,14,15,16]. In addition, the adoption of practices that improve soil carbon is also influenced by technological, cultural, production environment and the existing policy in a dissimilar way in different regions of sub-Saharan Africa [17, 18].

The adoption of innovations has been studied extensively in agriculture since the pioneering work on the adoption of hybrid corn in the United States of America (USA) by Griliches [19]. The attention of most studies on adoption was focused on finding answers to questions such as (1) what determines whether a particular farmer adopts or rejects an agricultural innovation, and (2) what exactly determines the diffusion pattern of agricultural innovation in a given population of potential adopters [20,21,22,23,24,25,26]. Despite many studies in adoptions, the findings from the majority of the studies often show contradiction as it relates to the importance and effect of any given variable on adoption. The contradiction can be accounted for by the dynamic nature of decisions related to the adoptions by farmers that are also conditioned by their perceptions and information as information continues to become available to them. This study, therefore, conceptualizes the adoption of SCEPs as a multistage decision process involving the acquisition of information and learning-by-doing from other farmers. This study also takes cognisance of what the theory of adoption refers to as entitlement and command over resources (i.e. natural, physical, human, financial and social capital) that the literature shows to have an effect on the adoption of innovations [27,28,29]. This is because agricultural innovations are usually introduced in a package to deal with a whole range of agricultural production constraints including low soil fertility and low soil carbon [30].

Despite the enormous literature on adoption of SCEPs, information on the factor that constrains or facilitates the adoption in eastern Africa is limited. The objective of this study is to systematically evaluate the current evidence base to identify: (a) which factors enable or constrain adoption of SCEPs and hence maintain soil carbon in Kenya and Ethiopia; and (b) to be able to lessons learnt concerning what influences the adoption of the SCEPs for the purpose of maintaining soil fertility among smallholder farms and how this can be improved going into the future for the purpose of formulating appropriate policies in both the short and long run. This can help us know whether the effect of specific socio-economic factors can be generalized during the formulation of policies aimed at effectively promoting and facilitating the adoption of SCEPs.

This review is organized as follows: Given the widespread definitions of carbon sequestration, it is vital to specify what this term and others associated with it means in the context of this review. Therefore, in Sect. 2, we provide a definition of the main terms and concepts used throughout this review. Section 3 describes the materials and methods used in this review and introduces the countries of focus. The results of this review are summarized in Sect. 4, while Sect. 5 highlights the conclusion and recommendations based on the findings from the review.

2 Methodology

2.1 Conceptual framework

In this study, we focus on the assessment of the factors that facilitate or constrain the adoption of SCEPs that we conceptually identify as both community (i.e. farmers group) and households level variables (i.e. socio-economic, fixed assets and institutional characteristics). Firstly, household endowment variables have been shown to influence or constrain the adoption or investment on SCEPs (Fig. 1). Abundant adoption literature has identified factors (such as age, education, gender, household size and incomes) that may influence the adoption process. The model developed by Feder et al. [22] presents one of the initial attempts to deal with interrelations in the adoption of agricultural innovation. The availability of labour for example—represented by the household size—is likely to influence the gross margin of innovation via its effect on yield. Income may influence the initial scale of adopting a given technology by relaxing financial constraints. Farmers’ age may influence risk aversion with the view that older farmers are likely to more risk-averse, and therefore likely to have a low adoption rate for new practices. Secondly, the effects of fixed assets such as farm and plot-level variables have also been shown to affect the adoption of SCEPs. The literature, however, underscores that the fixed assets must be borne in advance to allow the adoption of agricultural innovation over a given period. Consequently, inability to access a sizeable farm may interact with the dynamics of adoption (i.e. whether adoption of a given technology will pay-off over a given period) to bring about uncertainty, and this may lead to postponing of adoption. Thirdly, the adoption of SCEPs is also influenced by infrastructure and services variables such as access to credit, information, extension services and market and community group membership. Distance to the market, for example, acts as a proxy for transaction costs associated with the acquisition of inputs and marketing output. Access to the market can, therefore, affect the adoption of SCEPs [31]. In this review, the reader should take into account that none of the reviewed paper shows the general trends of the SCEPs. To bridge this gap, the present review attempts to provide a general overview of the effect of socio-economic factors on the different SCEPs adopted in Kenya and Ethiopia.

Fig. 1
figure 1

A conceptual illustration of the setup for systematic quantitative literature review

Full size image

2.2 Search strategy, context and study selection

In order to reduce the scope and challenges associated with the literature review, the literature review contained in this review was limited to only two countries—Kenya and Ethiopia—in eastern Africa. Kenya and Ethiopia have agro-ecological zones ranging from arid, humid and highlands, that experience a large variation in temperature, rainfall and vegetation cover—all of which determine the soil type and farming activities in these areas. Kenya and Ethiopia were chosen as the countries of focus based on the fact that this work is a part of a larger project that aims at scaling up the adoption in the two countries by 2021 [32]. Climate variability has become a threat to agricultural production—which majority (more than 80%) of the population in the both Kenya and Ethiopia depend on for their livelihoods. In addition, the farming systems in the two countries have become vulnerable and have weaker resilience capacity. Therefore, there is a need to invest in SCEPs, particularly those that have the potential (in both the short and long run) to improve crops and livestock yields—both of which are critical for wealth creation, food security and poverty reduction.

To be able to provide an answer to our research objective, we did a systematic analysis of the content of the peer-reviewed and published articles, technical reports and working papers following Pickering and Byrne [33]. The protocol of the ‘Preferred Reporting Items for Systematic Recommendations’ (PRISMA Statement) from Moher et al. [34] was used (Fig. 2). This approach has been applied widely in published systematic review literature in various subjects such as climate change, energy policy and justice [35], impacts of nature recreation on birds [36], impacts of infrastructure on vegetation and soil [37], failure for environmental sustainability policy implementation [38] and analysis of approaches for ecological restoration [39].

Fig. 2
figure 2

PRISMA statement describing the steps for systematic quantitative literature review

Full size image

A search for peer-reviewed literature was conducted in English language using the Google Scholar and Scopus online databases and the Boolean search function of “‘adoption’ OR ‘Willingness to pay’ OR ‘willingness to accept’ AND ‘Sustainable Land Management practices (SLM)’ OR ‘land management’ OR ‘agricultural practices’, AND ‘socio-economic factors’ OR ‘constraint’, OR ‘affect’ OR ‘facilitate’ OR ‘determine’ AND ‘investment’ OR ‘economics of Sustainable Land Management’ AND ‘Soil carbon’” between August and October 2017 to identify articles. From an initial search of 1020 documentation, 501 documents were excluded because they contained only one of the search term without any relation to the adoption of SCEPs, the first 519 documentations were assessed for publication type before dismissing documents that were duplicates, book chapters, reviews or grey literature (n = 218). The remaining 303 articles were further screened for relevance to the research objective, and 216 more articles were excluded as irrelevant. Additional 43 documents were excluded for quantitative analysis on the basis that they were either not reporting original data, discussed soil carbon but did not contain quantitative data. Consequently, a total of 44 articles were selected for detailed assessment (Fig. 2), and a total of 113 articles are cited in this paper.

The assessment criteria for the articles include the journal and authors; research methods used; subject countries; and the characteristics, influences and reported the outcome of factors constraining or enabling the sequestration of soil carbon (Table 1). We used the assigned field of research categories adapted from the Excellence in Research for Australia (ERA) 2015 submitted journal list [40]. In cases where the journal for the review was missing, the classification was done by the authors based on the publisher’s description and the scope of the journal.

Table 1 Criteria used to review the selected articles on what constrain or enable the adoption of practices that sequester soil carbon
Full size table

The assessment criteria for eligibility were done by reflecting on the mixed methods, after which we determined that qualitative, quantitative and policy/case studies were all eligible, provided that these studies (1) reported experience in one of the SCEPs, (2) included empirically derived information on determinants of uptake of any one of the SCEPs and (3) did not address a switch to other types of soil and land management practices that are not considered as SCEPs. We only considered studies if they had been conducted in Kenya and Ethiopia. We excluded studies undertaken in refugee camps due to the limited generalizability, along with studies that relied solely on the effectiveness of non-governmental organization rather than adoption and use by households.

As for the study selection, we employed a standardized approach in selecting the study with independent verification by at least two authors. The screening of title and abstract was done first followed by an independent random verification for 15% sample of included and excluded abstracts. Finally, the screening of full-text articles. To construct the data tables, the quantitative and qualitative data involving socio-economic factors that determine the adoption of SCEPs available from each of the reviewed study were done using piloted design-specific data extraction forms and summarized these in tables. The verification of extracted data was conducted during the synthesis. The methodological quality of each paper considered was scrutinized. For qualitative studies, Harden et al. [41] criteria cover quality of reporting, reliability of the data, the depth and breadth of analysis, and how well the results reflect the perspective of the studied households was applied. For the quantitative studies, the abridged version of the Liverpool University Quality Assessment Tools (LQAT), developed for prior systematic reviews (e.g. [42]), was used. These tools comprise several modules: sampling, experience, outcome, analysis, effects, assessment of bias and the overall classification of quality [43]. We organized the finding from this systematic review under four characteristics domains: household endowment, farm-level characteristics, infrastructure and services and biophysical characteristic (Table 2).

Table 2 The characteristic domains of factors whose influence was reviewed
Full size table

3 Data extraction and analysis

Following the conceptual framework (Fig. 1), the data extraction was done in alignment with the four domains, household endowment, farm-level characteristics, infrastructure and services and biophysical characteristic (Table 2). These domains were supported by descriptions and/or definitions from the included studies. The key concepts per domain were then used in the development of data extraction forms, while basic information from the reviewed studies and specific information related to the study objective formed the basis on which the data were extracted as study title, first author name, the year of publication, research design, country of study, study participants and/or stakeholders, interventions (i.e. SCEPs) level of implementation and implementation outcomes (i.e. effects), whether the outcome is positive or negative and the level of significance for each outcome for the quantitative studies and direction of influence for qualitative studies. Finally, the data were extracted for the four domains using a thematic analysis of the models contained in the reviewed papers to identify commonalities and differences. The extracted data were done independently by two reviewers for the selected studies and then jointly agreed upon by all review team. Once all the data were mapped deductively to the four domains, a separate inductive process of domains analysis was used to accommodate the qualitative data. One reviewer would extract the data and perform analysis by the four domains, with a second reviewer validating the results by independently extracting data and performing the analysis for a random sample of the reviewed papers. Finally, the results were then discussed with all members of the review team. Any disagreements were resolved through group discussion and consensus within the review team. This was then followed by the classification of the effect of the different factors in the four domains based on their final impact or effect on the adoption as a positive (+) or a negative (–) according to the reviewed studies. Before summarizing and analysing the explanatory variable affecting the adoption of SCEPs, we first present the description of each explanatory variable (Table 3). A database containing a detailed examination of the effects of the factors that constrain (i.e. negatively affect) or facilitate (i.e. positively affect) the adoption of technologies that increase soil carbon was first compiled in Microsoft Excel and then summarized in Microsoft Word.

Table 3 Variables and how they were measured in the reviewed articles
Full size table

4 Results and discussion

4.1 Nature of reviewed literature

About 40%, 24%, 16% and 8% of the quantitative articles had used Probit, Tobit and Logit models, respectively, in identifying socio-economic factors that enable or constrain the adoption of SCEPs among farmers. The remaining 6%, 4%, 2% and 1% of the reviewed literature had used mixed research methods, random effects models, optimization modelling and endogenous switching model, respectively. In Ethiopia, the average sample size was approximately 350 households with a standard deviation of 500 and more than 60% of the studies had been conducted in the Tigray and Amhara regions in the North and north-western part of Ethiopia, respectively. The remaining 40% were conducted in Oromia, Southern Nation Nationalities (SNNP) and Mixed region. Both Tigray and Amhara regions are situated in a highland and are characterized by a rugged landscape and are therefore prone to land degradation and soil erosion (http://www.idp-uk.org).

In Kenya, about 71%, 13%, 6% and 5% of the papers reviewed had used Probit, Probit, OLS and Logit models, respectively, in identifying socio-economic factors influencing the adoption of SCEPs. The rest (i.e. 5%) of the studies had used analysis of variance and partial budget. The average sample size of the households in the reviewed papers was 948 and a standard deviation of 2357.

In Kenya, the review showed that SCEPs are being used in varying intensities on different regions across the country (i.e. Nairobi, Coast, Western/Nyanza, Central, Eastern and Rift valley regions), and this is largely determined by the farming systems. For example, in arid and semi-arid regions (e.g. Eastern, Coast, and lower Central) where rainfall is low and/or unreliable, soils have poor water retention capacity and soil erosion is common leading to low soil organic matter in the soil. In these areas, common practices include soil bunds, fertilizer application, use of cover crops, multi-purpose trees and grasses [44, 45]. In high potential areas (e.g. Central Highlands, parts of Rift valley and Western) with well-drained fertile soils and reliable rainfall, the population pressure is a challenge. Consequently, intensive agricultural production is very common. In the Central Province, Rift valley and Western regions, soil nutrient replenishment measures such as the application of fertilizer and application of manure are common [46, 47]. This shows widespread use of SCEPs and thus a positive step maintaining soil fertility and therefore land sustainability. It also shows that households are willing to employ various techniques that have the potential of improving the quality of their land.

4.2 A review of determinants of adoption of practices that enhance soil carbon

The results from both Kenya and Ethiopia suggest that different socio-economic factors highlighted (Table 2) had a positive, negative or mixed effect on the adoption of SCEPs (Table 6).

4.2.1 Household endowment

Age of the household head (Age_hh) The age of household head positively influences the adoption of the cut-off drain, cover crops and multiple purpose trees, while it had a mixed effect on the adoption of crop residues, soil/stone bunds in Ethiopia (Tables 4, 6). In Kenya, the age of household head had a positive effect on the adoption of minimum tillage and erosion control (Tables 5, 6). The positive influence of the household age on the adoption of these practices suggests that they require less labour and more experience to install over time. Their maintenance also is less labour demanding. However, the age of the household head constrains the adoption of conservation tillage, drainage ditches, fanya juu terraces, grass strips, gully traps, inorganic fertilizer, intercropping, minimum tillage and soil water conservation in Ethiopia (Table 4), and crop residues, intercropping and inorganic fertilizer in Kenya (Table 5). This could be due to constraints in terms of labour and financial resources and access to information and inputs facing aged household heads [48, 49]. Practices that require large cash outlay (i.e. inorganic fertilizer) are perceived as risky among the aged household heads compared to young household heads [48, 50]. The effect of age of the household on the adoption of SCEPs is therefore not uniform [51].

Table 4 Influences of different factors on the adoption of soil carbon enhancing practices in Ethiopia
Full size table
Table 5 Influences of different factors on the adoption of soil carbon enhancing practices in Kenya
Full size table

Gender of the household head (Gend_hh) Male-headed households are more likely to adopt conservation of tillage, crop residues, ‘fanya juu’ terraces, farmyard manure, cover crops, soil and water conservation and the multi-purpose trees in Ethiopia (Tables 4, 6). In Kenya, male-headed households are more likely to adopt ‘fanya juu’ terraces, crop residues, farmyard manure, intercropping and soil/sand bunds in Kenya (Tables 5, 6). This could be because these practices are labour demanding to establish, and male-headed households are less constrained in terms of resources and can, therefore, mobilize extra labour easily compared to their female counterparts [48, 84, 86]. In their study, Ndiritu et al. [82] show that male-headed household easily applies animal manure, as they can employ additional labour compared to female-headed households in Eastern and Western Kenya.

In Ethiopia, male-headed households are less likely to adopt farmyard manure in Ethiopia (Table 4) and minimum tillage in Kenya (Tables 5, 6) suggesting that perception about the usefulness of specific practices in terms of improving soil fertility and enhancing yield among decision-makers could be playing a critical role [51]. This observation is sustained by García de Jalón et al. [91] who observed that male-headed households in Makueni in Kenya have a sceptical response to climate change which acts as a behavioural barrier to adoption. However, the effect of gender on the adoption of minimum tillage and soil/stone bunds was mixed for Ethiopia.

Education of the household head (Educ_hh) Educated household heads were more likely to adopt farmyard manure, ‘fanya juu’ terraces, grass strips, gully traps, inorganic fertilizers, intercropping, multi-purpose trees, soil/stone bunds and soil water conservation and cover crops in Ethiopia (Tables 4, 6), while they were more likely to adopt soil/stone bunds in Kenya (Tables 5, 6). The positive effect of education on these SCEPs could be due to the enhanced understanding of how to implement them [92] and an understanding of their importance in maintaining soil fertility [83, 85, 86]. In Ethiopia, education has been associated with increased investment in farming [73].

The observed negative influence of education on the adoption of crop residues and minimum tillage in Ethiopia (Tables 4, 6) and farmyard manure, intercropping and multi-purpose trees in Kenya (Tables 5, 6) could be because it helps the households to evaluate practices in terms of their cost and benefits so that practices that are less profitable are also less appealing for adoption. The negative influence of education could also be due to improved access to off-farm income-generating activities [10]. Nevertheless, the education of household head had a mixed effect on the adoption of crop rotation in Ethiopia (Table 6). These findings suggest improving the uptake of SCEPs among farmers need careful consideration of literacy levels among farmers.

Household size (Hh_size) In Ethiopia, household size had a positive influence on the adoption of all SCEPs studied, except the adoption of drainage ditches (Table 4), while in Kenya it had a positive effect on soil and water conservation and farmyard manure (Table 5) suggesting that implementing SCEPs could be labour demanding [60, 78, 82, 84, 93, 94]. Labour availability is, therefore, crucial for the adoption of SCEPs [55, 75, 82, 84, 93]. Nevertheless, household size had a mixed effect on the adoption of farmyard manure and inorganic fertilizer in Ethiopia and Kenya, respectively. The reason for this observation could be because the use of farmyard manure is a labour-intensive activity considering collection and application while fertilizer requires resources to acquire. Therefore, both of these practices can introduce as an additional constraint depending on the size of the households and their needs [83].

Off-farm income (Off_inc) Off-farm income had a positive effect on the adoption of crop residues, grass strips and ‘fanya juu’ terraces, in Ethiopia (Table 4), and soil and water conservation and crop residues in Kenya (Table 5), suggesting that off-farm income relaxes liquidity constraint thereby facilitating the adoption of practices that require cash outlays (including acquisition of labour and inputs) [95]. In Western Kenya for example, adoption of inorganic fertilizer is common among the wealthy farmers as compared to the poor farmers [47, 83, 96]. Off-farm income is therefore important especially in areas with poorly developed credit market [73]. However, off-farm income exerts a negative influence on the adoption of crop rotation, grass strips and intercropping in Ethiopia, crop residues and farmyard manure in Kenya (Tables 4 and 5), suggesting that engagement in other off-farm income-generating activities diverts labour required for implementing SCEPs [67, 85, 97]. Thus, there may be a need for prioritizing farm objectives that can lead to sustainable farm production and livelihood in the long-term.

Livestock holding (Hrd_sze) Herd size has a positive effect to the adoption of cut-off drain, multi-purpose trees, compost manure, ‘fanya juu’ terraces, farmyard manure and soil and water conservation in Ethiopia (Table 4), and intercropping and inorganic fertilizer in Kenya (Table 5) suggesting that livestock could be complementing the implementation of these practices through the income they generate and hence their positive influence [10, 98]. However, the effect of livestock holding on the adoption of crop residues in Ethiopia (Table 4) and soil and water management in Kenya (Table 5) suggesting that livestock could be consuming residues and other available resources that could be used in soil and water management [99].

4.2.2 Farm-level characteristics

Farm size (Frm_sze) Farm size had a positive influence on the adoption of all the soil enhancing practices studied in Ethiopia, except conservation agriculture, intercropping and minimum tillage (Table 4). In Kenya, farm size had a positive influence on the adoption of multi-purpose trees (Table 5), suggesting that these practices may not be strictly scale neutral or that the opportunity costs facing households vary by their farm size [86, 100]. The positive effect could also be farm size which has been shown to be correlated with wealth which relaxes financial constraint when used as collateral for the acquisition of credit [95]. Nevertheless, farm size had a mixed effect on the adoption of conservation tillage, crop residues, intercropping, soil/stone bunds and inorganic fertilizer, and a negative effect on minimum tillage in Kenya (Table 5). The negative and the mixed effect of farm size on the adoption of minimum tillage is because the farm size may influence farmers decision differently. For example, a farmer may not worry about soil erosion and degradation, hence reducing their willingness to adopt them [10, 53, 61, 74]. Small farm size may also hinder the adoption of practices that have the potential to sequester carbon [46, 78, 101,102,103].

Plot slope (Plt_slop) Plot slope had a positive effect on the adoption of all the studied SCEPs except contour ploughing, crop rotation, minimum tillage and compost manure in Ethiopia (Table 4), and ‘fanya juu’ terraces, farmyard manure and crop residues in Kenya (Table 5), suggesting that farmers are aware of the negative impact of slopes—such as erosion and loss of soil fertility—on soil carbon. Many soil erosion and fertility loss prevention measures are practised on sloping lands [74, 79, 82].

4.2.3 Infrastructure and services

Access to credit (Crdt_acss) Access to credit has a positive effect on the adoption of compost manure, crop residues, ‘fanya juu’ terraces, inorganic fertilizer, minimum tillage and soil/stone bunds in Ethiopia (Table 4), and inorganic fertilizer in Kenya (Table 5). Credit relaxes households binding financial constraints to acquire inputs such as fertilizer and probably hire labour. Minimum tillage is not so costly but is an effective measure for ensuring soil fertility [92, 104, 105]. In Kenya, access to credit has a negative effect on crop rotation, minimum tillage, soil and water conservation suggesting that decision relating to the adoption of SCEPs could also be influenced by other hidden issues such as the perception of the decision-makers and the environment among others.

Information access (Info_acss) The access to information has a positive effect on the adoption of inorganic fertilizer, soil and/or stone bunds in Kenya (Table 5), suggesting that information access provides the technical know-how required for the implementation of these practices. Information access can also lead to adoption because of the realization of the benefits associated with the adoption of these practices [86, 87, 89, 103, 106]. Bridging the gap through the provision of benefits associated with the adoption of specific practices increases the adoption or rate of diffusion [81, 82, 107]. Information access, however, tends to constrain the adoption of intercropping (Table 5) suggesting that information access tends to influence the adoption decision for example through comparing the costs and benefits of a practice [108].

Access to the market (Dist_mkt) Distance to the market has a positive effect on the adoption of crop residues, crop rotation, farmyard manure and minimum tillage in Ethiopia (Table 4), and crop rotation and minimum tillage in Kenya (Table 5) suggesting that high transaction may influence the adoption of technologies that rely less on the market to implement and maintain. Distance to the market had a mixed effect on the adoption of the multi-purpose trees, gully traps and soil/stone bunds in Ethiopia (Table 4), and inorganic fertilizer and intercropping in Kenya (Table 5) suggesting that the adoption of SCEPs depends on other factors beside socio-economic factors. For example, Murage et al. [88] show that improved access to the market is associated with the adoption of inorganic fertilizer, while poor market access tends to constrain it [84]. This finding suggests that the adoption of SCEPs practices could also be due to other factors beside socio-economic factors, and therefore, the scaling up of these practices need not be based on the finding from socio-economic discipline only. There is a need for a multidisciplinary approach to help understand what besides the socio-economic characteristics influences the adoption of SCEPs.

Farmers’ organization (Grp_memb) Group memberships have a positive effect on the adoption of conservation tillage, crop rotation, fanya juu terraces, farmyard manure, grass strips, intercropping, multi-purpose trees, soil/stone bunds and soil and water conservation in Ethiopia (Table 4), and ‘fanya juu’ terraces, minimum tillage, soil/stone bunds and intercropping in Kenya (Table 5). Farmers groups are a source of information exchange, capacity building and a form of social capital [98], suggesting that farmers are able to access information about the benefits of different technologies leading to their adoption [84, 106].

Extension services (Ext_serv) Extension services have a positive effect on the adoption of all the SCEPs studied in Ethiopia, except contour planting, cut-off drains, drainage ditches, grass strips, cover crops and gully traps (Table 4). In Kenya, extension services exert a positive influence on the adoption of soil and water conservation, farmyard manure, minimum tillage, inorganic fertilizer, crop residues and intercropping (Table 5). This suggests that extension services play a crucial role in enhancing learning which leads to the adoption of new technologies [103]. Lack of extension services, therefore, can act as an impediment to the adoption of soil carbon enhancement practices. However, since the extension services do not affect all the practices in a uniform way, policy formulation should take this into account, by not generalizing that it affects the adoption of SCEPs positively.

4.2.4 Institutional characteristics

Land tenure (Lnd_tnr) Land tenure had a positive effect on the adoption of inorganic fertilizer, soil and water conservation, compost manure and soil and water conservation in Ethiopia (Table 4), and farmyard manure and stone and/or soil bunds (Table 5), suggesting that tenure security provides an impetus to farmers for implementing long-term SCEPs [79, 86]. However, tenure security has a negative effect on the adoption of farmyard manure, grass strips, intercropping, multi-purpose trees in Ethiopia (Table 4), and erosion control and inorganic fertilizer in Kenya (Table 5). Technologies that demand high labour and capital to implement results in spiralling expenditure and given the difficulty in getting sufficient income for employing labourers, they are prohibitive [55]. The mixed effect of land tenure on soil/stone bunds could be due to the differences in the decision-making process as influenced by the type of land ownership (i.e. whether the land is rented or owned) (Ibid). However, farmers who own lands could use their title deeds as collateral to access credit.

4.3 Summary of the review

For ease of identification of the distinct direction of effects of the various SCEPs reviewed, Table 6 provides a general overview of trends showing the influence for the socio-economic factors on SCEPs in Kenya and Ethiopia. This was achieved by qualitatively summarizing the direction of socio-economic, institutional and biophysical influences reported by all reviewed articles on different SCEPs reported in Tables 4 and 5. The factors that constrain the adoption of SCEPs are labelled ‘negative’ while those that facilitate are labelled ‘positive’ on columns three and four. Those that have a mixed effect, i.e. the positive and negative effect is labelled ‘mixed’. This summary is easier to understand by policymakers, academic and non-academic audience. The results show that although most of the household and farm-level variables have either positive or negative on most of the SCEPs, some inconsistencies were observed suggesting that going forward scientists need to take care when making recommendations based on studies that are not multidisciplinary in nature. This is because, as the review shows, the decision to adopt SCEPs is definitely influenced by socio-economic factors, but the mixed effect by some of the socio-economic factors on some of the SCEPs suggests that there could be other factors such as the environment, history, culture and agro-ecology that could be playing a major role in the adoption of SCEPs, and these need to be considered before policy prescription or recommendation intended for scaling up SCEPs adoption is made.

5 Conclusion and recommendations

This paper provides a review from a large amount of quantitative and qualitative evidence derived from peer-reviewed articles on the socio-economic factors that influence the adoption of various SCEPs. The discussion focused on the influences of household endowment (age, education, gender, household size, livestock holding, farm size and off-farm income), farm-level characteristics (farm size and plot slope) and infrastructure and services (access to credit, access to information, access to extension services, access to the market and membership to farmer’s groups) and institutional characteristics (land tenure) on SCEPs. Thus, our review and synthesis underscore the need for taking a multidisciplinary approach in trying to understand why some practices exhibit different patterns of effect from the socio-economic factors.

The review further provides evidence that can help to evolve thinking and policy prescriptions for enhanced adoption of SCEPs. Significant factors that motivate farmers to adopt different SCEPs vary from the household endowment, farm-level characteristics, infrastructure and biophysical factors. The direction of influence by specific factors also tends to vary from positive, negative to mixed effect on various practices in the two countries. This is an indication that policy prescription need not be rushed if sustainability in scaling up the adoption of SCEPs is to be achieved. While most agricultural policies in East Africa delve on infrastructural development for upscaling adoption of SLMPs, this review reveals that policy interventions should target household endowment so as to improve farmers’ capacity to adopt SCEPs. For instance, the varied effects of age, education, information access and extension services on the adoption of SCEPs suggest innovative methods in the dissemination of information on these practices. Accommodating both the young and old age groups needs in information delivery, increasing the frequency of training, workshops, seminars and on-farm field demonstrations are feasible in creating awareness for suitable SCEPs in specific geographical spaces. From the review, the impact of off-farm income and access to credit in the adoption of SCEPs are clear-cut, thus approaches such as improving households’ access to off-farm income-generating opportunities and encouraging farmers to join informal rural credit schemes can be used to enhance adoption.

Similarly, our review highlights the effect of gender in the decision to adopt SCEPs. Generally, women compared to their male counterparts lack resources which are required to implement these practices. This is an implication that policy formulation relating to the adoption of SCEPs should be gender-sensitive. Efforts by researchers in coming up with improvised versions of SCEPs that are less labour-intensive and cost-effective are recommended in encouraging women engagement in the adoption of SCEPs. The significance of tenure security in the adoption of SCEPs cannot be overemphasized. This calls for government’s involvement in coming up with contracts as a way of formalizing tenure agreements among farmers, a step that would encourage them to make investments on long-term SCEPs such as terracing. Also, farmers could be encouraged to mobilize themselves in groups and acquire land as a mass other than individually investing in small segmented pieces of land that are not economically viable to invest in.

Important to note is that appropriate soil carbon enhancing practice may differ from one location to another depending on the agro-ecological, socio-economic and market conditions. Therefore, rather than searching for a general blueprint, the review also emphasises for looking at factors that tend to constrain or facilitate the adoption of different technologies in a different context, and as such open an avenue for identifying complementarities and trade-offs.

This study further counters the misconception that socio-economic factors affect the adoption of carbon enhancing technologies in either a strictly positive or negative way. This is because the importance of the different factors and the direction of influence vary depending on the nature of the practice in question. This indicates therefore that the adoption of SCEPs practices could be due to other factors beside socio-economic factors, and therefore, the scaling up of these practices needs not be based on the finding from socio-economic discipline only. Taking this into account before scaling up of SCEPs adoption will then allow the interrogation of the likely effect of given socio-economic characteristics on the adoption of specific SCEPs under varying environments, and to ask ourselves what else is missing that can help the adoption patterns and decision. As such, there is a need for further deeper multidisciplinary search of what else accounts for mixed effects of some of the households’ endowment, infrastructure and services, farm-level characteristics and biophysical characteristics on specific SCEPs in a given region and farming systems. Such an approach will provide more robust findings on which targeted recommendations and economic incentives for improved adoption of different soil enhancing practices in different locations can be based. There is a need for a multidisciplinary approach to help understand what besides the socio-economic characteristics influences the adoption of SCEPs. Moreover, such findings may help to design trans-generation policies aimed at increasing the development and uptake of SCEPs in East Africa. Adoption of SCEPs informed by such robust information is likely to be more sustainable in the long-term.