Understanding the contexts of effectiveness of adaptation to climate change and variability: a qualitative study of smallholder farmers in central Ethiopia

ABSTRACT

Although most studies showed the positive impact of climate change adaptation on farmers’ livelihoods, not all strategies are effective. Based on a qualitative study of smallholder farmers in central Ethiopia, this study examined the roles of contextual factors in contributing to or constraining the effectiveness of adaptation strategies. The results show that (i) strategies heavily reliant on rainfall were less effective; (ii) farmers with fatalistic attitudes utilized survival-oriented and low-return strategies, which were less effective to protect them from the effects of climate change and variability (CCV); (iii) farmers who had access to livelihood assets effectively responded to CCV through diversification of adaptation strategies and involvement in high-return strategies; (iv) Inadequate extension services, inefficient supply of farm inputs, and limited inclusivity of poor farmers undermined the roles of formal institutions in supporting effective adaptation. The findings attest to the importance of thoroughly understanding the wider contexts in which adaptation strategies are implemented to ensure their effectiveness. Building the adaptive capacity of farmers through pro-poor measures opens opportunities for engagement in high-return adaptation strategies, thereby promoting the building of sustainable livelihoods by all farmers.

KEYWORDS:

• Livelihood
• adaptive capacity
• institution
• impact
• fuzzy cognitive map

1. Introduction

Smallholder farmers in Ethiopia are vulnerable to Climate Change and Variability (CCV) due to excessive dependence on agriculture in a fragile environment. Hence, adaptation to CCV has become a priority to improve their livelihoods. In addition to strategies implemented by various organizations, farmers autonomously adapt to observed or predicted rainfall variability and extreme weather events. These include, among others, changing planting time, adjusting cropping practices, diversifying crops, planting drought-tolerant improved seeds, using irrigation, undertaking land management activities, involving in non-farm activities, and migrating to other areas (Etana et al., 2020a). These adaptation decisions yield sustainable livelihood outcomes if they ensure synergies between economic, social, and environmental benefits (Haggar et al., 2021).

Most studies in Ethiopia (Di Falco & Chavas, 2009; Di Falco et al., 2011; Teklewold et al., 2013) showed the positive impact of adaptation on food security, income, and crop yields. However, adaptation strategies are not always effective (Eriksen et al., 2011). First, adaptation might be ineffective in protecting livelihoods from the effects of CCV. For instance, migrants may not get jobs at their destination, which makes lives more difficult for them and for the sending households. The reduction of farm labour due to migration also affects farm operations, resulting in increased vulnerability (Antwi-Agyei et al., 2018). Second, adaptation may be effective in the short term, but may increase vulnerability in the long term (Moser & Ekstrom, 2010). For example, irrigation may turn out to be maladaptive due to climate-induced water scarcity (Antwi-Agyei et al., 2018). Third, not all households equally benefit from the use of adaptation strategies (Douxchamps et al., 2016). Fourth, adaptation may create unintended outcomes, and socio-economic and environmental problems (Eriksen et al., 2011). Fifth, there are also trade-offs between different outcomes of farmers’ decisions (Haggar et al., 2021).

A nuanced understanding of the variation of the impact of adaptation requires careful analysis of the contexts in which the strategies are used (Rodriguez-Solorzano, 2014). Establishing a statistical relationship between the use of adaptation strategies and livelihood outcomes is not sufficient to understand the reasons the strategies are effective or not. Scaling-up of adaptation strategies hinges on proper recognition of the contexts in which the intended objectives of adaptation are achieved or conversely the outcomes are negative or maladaptive (Juhola et al., 2016). These necessitate an in-depth understanding of the contexts in which adaptation strategies are implemented and the specific interactions between diverse contextual factors contributing to, or constraining, effective adaptation. However, there is a lack of evidence on the factors that influence the effectiveness of different adaptation strategies in counteracting the effects of CCV. Therefore, the objective of this study is to investigate the reasons adaptation strategies are (in)effective in improving the livelihoods of smallholder farmers. The study provides an empirical basis that will inform the design and implementation of contextually relevant adaptation strategies to build sustainable rural livelihoods.

2. A contextual approach to effective adaptation: conceptual and empirical accounts

There is no consensus on what constitutes effective adaptation. The differing and contested meanings given by different scholars challenge its definition and measurement (Eriksen et al., 2021). According to Adger et al. (2005), adaptation is considered effective if it can achieve its stated objectives. In line with the sustainable livelihoods framework (SLF), these objectives constitute economic efficiency, social equity, and environmental viability (Scoones, 2009). This entails that effective adaptation is responsive to various social, economic, and environmental contexts (Dessai & Hulme, 2007). These three criteria are interdependent as the benefits obtained in one dimension may be positively or negatively related to the outcomes in another dimension(Connolly-Boutin & Smit, 2016). Hence, the use of different sets of indicators is important to evaluate the effectiveness of adaptation.

Through its focus on conditions and processes, a contextual approach gives in-depth insights into the enablers or constraints of effective adaptation (O'Brien et al., 2007). A contextual approach involves analysis of the complex interactions between a wide range of factors contributing to or hindering the effectiveness of adaptation (Burnham & Ma, 2018). It reveals the multiple pathways to effective adaptation and the differential outcome of using adaptation strategies. In this vein, the SLF provides a people-centred approach to understanding the contexts in which smallholder farmers make a living (Scoones, 2009). The contextual factors influencing the effectiveness of adaptation can be categorized into climate and non-climate risks (biophysical factors), socio-cultural factors, livelihood assets, and institutional factors. These contextual factors, which are not mutually exclusive, result in differing livelihood outcomes of adaptation (Etana et al., 2021b; Moser & Ekstrom, 2010).

Socio-cultural contexts (e.g. values, norms, and risk perceptions) and adaptive capacity shape the process and effectiveness of adaptation (Adger et al., 2005; Moser & Ekstrom, 2010). Effectiveness is value-laden in that the objectives of adaptation, types of adaptation strategies, and the outcomes of the strategies are rooted in farmers’ values and norms of what they recognize to be worth achieving (O'Brien & Wolf, 2010). The effectiveness of adaptation is also situated in people’s perceptions of risks (Wyborn et al., 2015). When farmers perceive that climate change is less significant to their livelihoods than non-climate-related events, adaptation interventions are less effective in offsetting the effects of CCV (Patt & Schroter, 2008). Besides, when farmers perceive that they have no capacity, they either fail to take action or make very little effort in using adaptation strategies. Adaptive capacity is also at the core of effective adaptation. It constitutes, according to the SLF, five types of assets: natural, physical, human, financial, and social (Scoones, 2009). Access to these assets determines the type, number, and effectiveness of adaptation strategies (Asfaw et al., 2016; Jain et al., 2015; Nyberg et al., 2021). Even the same adaptation strategies are effective at varying degrees due to differences in households’ adaptive capacity (Douxchamps et al., 2016).

The process and outcome of adaptation are the functions of the larger contexts (i.e. institutional, climatic, and non-climatic) in which the strategies are used (Moser & Ekstrom, 2010). Formal institutions support effective adaptation by shaping farmers’ perceptions, counteracting maladaptive practices through an institutional framework that fosters learning; supplies of efficient technologies; and promotion of the use of more effective adaptation strategies (Asfaw et al., 2016; Islam & Nursey-Bray, 2017; Nyberg et al., 2021). Institutions can also build adaptive capacity by facilitating access to resources and speeding up recovery from the adverse effects of CCV (Yaro et al., 2015). Governance issues such as inclusion and participation as well as accountability are important in shaping the adaptation decisions of smallholder farmers (Agrawal et al., 2008). Informal networks play key roles in supporting farmers’ adaptation efforts by expediting their access to information and farm inputs (Asfaw et al., 2016). Market factors are also among the institutional drivers of effective adaptation (Burnham & Ma, 2018; Matz et al., 2015; Yaro et al., 2015). The extent of effectiveness of adaptation strategies varies under different climate and non-climate contexts. For instance, farm-level adaptation strategies (e.g. using improved seeds and fertilizers) improve economic returns when there is optimum temperature and precipitation (Arslan et al., 2015; Waha et al., 2013). Non-climate factors also shape the effectiveness of adaptation strategies (Douxchamps et al., 2016; Moser & Ekstrom, 2010; Rodriguez-Solorzano, 2014). Hence, a holistic inter-disciplinary framework integrating socio-cultural, economic, institutional, and biophysical factors is utilized in this study to investigate the effectiveness of adaptation strategies.

3. Data and methods

3.1. Description of the study area

This study was conducted in three districts (Kembibit, Kuyu, and Boset) in central Ethiopia. The selection was made based on the representation of different agro-ecological settings, the similarity of livelihood systems (i.e. mixed farming), vulnerability to CCV, and prevalence of food insecurity. The three districts dominantly represent highland (H), midland (M), and lowland (L) agro-ecological settings, respectively. The average annual temperature of the three areas is 14.6, 15.5, and 22.1⁰C, respectively (Etana et al., 2020b). The areas are characterized by a bimodal rainfall distribution with a short belg rainy season between March and May, and a long kiremt rainy season between June and September. The long-term average rainfall of the short rainy season was 72, 177, and 81 mm in the highland, midland, and lowland areas, respectively. The long-term average rainfall of the long rainy season was 690, 944, and 384 mm in the same order (Etana et al., 2020b). The livelihood of the population depends on crop and livestock production. Farmers follow the bimodal rainfall distribution to produce crops. There is a high risk of yield reduction or crop failure during years of adverse weather conditions. In addition to rainfall variability, extreme events such as drought, flood, frost, and snowfall occur, to varying degrees, in the study area. Due to exposure to recurrent climate risks, a sizeable proportion of the population is under the support of the Productive Safety-net Program aimed at reducing vulnerability to CCV and building livelihoods.

3.2. Sources of data and methods of data collection

A qualitative research method was used to examine the contextual factors influencing the effectiveness of adaptation strategies. Specifically, face-to-face focus group discussions (FGDs), in-depth interviews (INDs), and key informant interviews (KIIs) were used to collect the data. The sources of data were male and female household heads, development agents, disaster risk management officers, agricultural extension officers, food security officers, kebele administrators, and representatives of microfinance institutions and non-governmental organizations who were selected through purposive sampling. A total of 12 FGDs, 30 INDs, and 19 KIIs were conducted in private places around their residential area or workplaces and at a time convenient for them. There was no refusal to participate in the study. The average time required was one hour and one and half an hour for interviews and focus group discussions, respectively. Checklists with a set of semi-structured questions were used to guide the discussions and interviews. The guiding questions mainly emphasized the livelihood problems of farmers, the types of adaptation strategies they use, the roles of the strategies in protecting them from the effects of CCV, the reasons for the effectiveness or otherwise of different adaptation strategies, and the roles of institutions in supporting effective adaptation. The data were collected following the approval of the fieldwork activities by the pertinent government offices at different levels of the administrative structure and obtaining verbal informed consent from the research participants. After obtaining the oral consent of the participants, the discussions were audio-recorded.

3.3. Data analysis

The audio-recorded discussions and interviews were transcribed and translated. The translated document was analysed inductively to systematically organize the data into a structured format and a coherent line of argument. Then, they were analysed using a thematic approach and fuzzy cognitive mapping. The thematic analysis, which offers detailed and rich descriptions of data (Braun & Clarke, 2006), was undertaken in a series of steps. First, themes were identified based on the research objective. Second, open coding schemes were established through repeated reading of selected transcripts. Third, the transcripts were coded following the coding scheme. Then, similarities, differences, and connections between the contents in the coding scheme were analysed to get a qualitative understanding of the data. ATLAS.ti software was employed for coding and analysis.

The fuzzy cognitive map (FCM) was used to graphically depict the causal relationships between variables influencing the effectiveness of adaptation (Gray et al., 2013). FCM is a useful tool to analyse a complex adaptation system based on a qualitative study by integrating a broad range of knowledge obtained from stakeholders (Özesmi & Özesmi, 2004), which allows for a deeper understanding of the contexts of adaptation effectiveness. Unlike other models that require large datasets, that are either unavailable or costly to collect, FCM models human behaviour based on a small number of observations (Özesmi & Özesmi, 2004). The social cognitive map not only shows how individuals or groups internally construct their understanding of how they adapt to CCV but also represents shared knowledge in a community (Gray et al., 2013; Gray et al., 2014).

The map was constructed following a series of steps. First, the stakeholders were identified to generate data. Accordingly, data were collected from farmers (30), development agents (3), local administrators (3), agricultural extension (3), food security (3), and disaster risk management officers (3), and one NGO staff from the three study districts. Second, variables that describe the behaviour of the smallholder farmers’ livelihood adaptation system were identified. Third, causal relationships were established between the variables using directed arrows showing either positive or negative relationships. Fourth, the magnitude of the relationship was determined based on the study participants’ normative judgement. In the fourth stage, the connections of the FCM model were represented in a matrix form from which several types of variables were identified (Gray et al., 2014). These were: driver variables (they influence other variables but are not themselves influenced by other variables), receiver variables (they are influenced by other variables but do not influence any other variable), and ordinary variables (they influence other variables and are also influenced by other variables). Finally, the individual cognitive maps were aggregated into a shared cognitive map by condensing the variables into broader categories to show the complex relationship between the process of adaptation and its outcomes (Harary et al., 1965). The relationships between the variables of the FCM were shown diagrammatically using the modelling tool called ‘Mental Modeller’ (Gray et al., 2013). The entire process of study design, data collection, data analysis, and report writing was guided by the Consolidated Criteria for Reporting Qualitative Research (see the supplementary file for the details of the codes and themes, properties of the cognitive maps, and the overall study process). Through researchers’ reflexivity, an iterative approach was used to link data, analysis, and findings.

4. Results

4.1. Description of the sample

A total of 113 individuals have participated in the 12 focus group discussions. The majority of them (75%) were males. Individuals in the age group 30–59 constituted the largest proportion of the FGD participants, followed by those in the age group 60 and older (20%). More than half of the participants (53%) had no education and 43% had primary-level education. In-depth interviews were conducted with 9 and 21 female and male farmers, respectively. Three agricultural extension, food security, and disaster risk management officers as well as development agents and local administrators participated in the key informant interviews. In addition, they were conducted with two staff of microfinance institutions, one staff of an international NGO, and one staff of an agricultural research center.

4.2. Definition of effectiveness

The effectiveness of adaptation strategies was conceptualized differently. Farmers’ most widely used criterion was food security, described as feeding a family for the whole year from their production. Farmers perceived that those effectively adapting to CCV do not rely on the market to buy grain or seeds. Effectiveness also involves the generation of additional income from diverse sources. It was also defined in terms of asset building. This includes productive assets such as oxen, hybrid cows, a ‘large size’ of farmland, and non-productive assets such as houses with corrugated-iron roofs, television, and sofa seats. Livelihood improvement was characterized by households’ savings to be used to buy farm inputs and other consumption goods, cover household expenses, and purchase grain during years of adverse weather conditions. In the discussions held with local government officers and NGOs, effectiveness was similarly conceived as increasing crop production, ensuring food self-sufficiency, raising household income, building household assets, maintaining the productivity of farmland, and improving the living standard of households.

4.3. Adaptation outcomes

Farmers underscored that adaptation had several positive outcomes. The most anticipated benefit was the improvement in the economic well-being of households. For instance, the positive outcome of adjustment of cropping practices, irrigation, and land management activities was an increase in crop production. Through the use of improved dairy cows and irrigation-based production of fruits and vegetables, farmers increased their production which was used for household consumption and generation of additional income. Adaptation also helped farmers to build assets including livestock and household equipment. Denoting the environmental benefits of adaptation, terracing and water-shed management activities contributed to the reduction of erosion, improvement of soil fertility, and restoration of forest coverage in some pocket areas.

However, some adaptation strategies turned out to be maladaptive or increased vulnerability. Out-migration, for instance, reduced households’ labour availability, thereby undermining local livelihood activities. Farmers who had limited adaptation options resorted to maladaptive responses such as deforestation and the production of charcoal, which affects long-term livelihood resilience. For some households, children failed to attend school due to parents’ lack of capacity to afford their schooling or due to engagement in income-generating activities such as daily labour work, which has lasting negative consequences on their livelihoods. Failure to increase production due to rainfall variability also induced the use of strategies that adversely affect farmers’ livelihoods in the short or long term. When there is crop failure or yield reduction, farmers strive to survive through asset disposal (e.g. selling oxen), which, however, deteriorates livelihoods in the long term. Limited livelihood options and failure of adaptation strategies also induced feelings of desperation, particularly among poor farmers.

4.4. Perceived effectiveness of the adaptation strategies

Livestock production (animal fattening, rearing dairy cows, and poultry) was among the strategies perceived by farmers to be effective. Rearing dairy cows and income earned from the selling of milk products supports farmers in withstanding the problem of CCV. As noted in one of the discussions, “Farmers earn up to 9000 birr [equivalent to 166 Euros] per month through the sale of milk which is greater than what most government employees earn.” [FGD-M-7]. Households using diversified strategies were also able to effectively adapt to CCV. It helped to generate a “good amount” of money that would be used to cover the costs of family members’ clothing, children’s schooling, and other household expenses. Irrigation was an effective strategy in the lowland areas. Farmers noted that “the life of many people is becoming better since they have started using irrigation.” [FGD-L-12]. It is used to produce market-oriented products (fruits and vegetables). It also allowed production multiple times a year. In the highland and midland areas, however, it was moderately effective due to limited access to water. The use of improved seeds [drought-tolerant and fast-maturing seeds] played a crucial role in livelihood improvement. It was indicated that “farmers using improved seeds benefit more than those who are not using in terms of getting a higher yield.” [FGD-L-10]. Likewise, land management activities helped in the retention of water and moisture thereby reducing crop loss and increasing yields.

The adjustment of planting time, switching crop type, migration, and non-farm activities were less effective in all settings. Changing planting time was less effective due to variability in the time of onset and cessation of rainfall.

“When we assume that the rain will likely terminate early, we plant earlier than the usual time. Unfortunately, this is not successful as the soil is too dry to allow the growth of seeds. I tried this, but it is not effective.” [IND-M-23].

Switching crops partly involved abandoning of production of some crops, which reduced options for crop diversification. “There is no belg rainfall now. Farmers also refrained from sowing belg crops.” [FGD-H-1]. This quote further implies that failure to produce belg crops results in the reduction of the total amount of yields to be obtained. Farmers also underscored that migration was used only as a means of survival by resource-poor households due to a lack of lucrative job opportunities and high costs of living at the destinations. Limited non-farm job opportunities, engagement in low-paying activities, and increasing prices of grain bought from the market for household consumption made non-farm activities less effective.

The differing outcomes of adaptation are the function of the contexts in which the strategies are pursued. The contextual factors shaping the pathways of effective (Figure 1) and ineffective (Figure 2) adaptation are drawn using FCM. In these figures, variables referring to climate and non-climate risks were represented by light orange colour. Light yellow and light purple colours were used to distinguish variables representing socio-cultural factors and livelihood assets, respectively. The characteristics of adaptation strategies were represented by light aquamarine colour while light blue colour was used to denote variables showing institutional factors. The silver colour shows the outcome variable (i.e. effectiveness).

Figure 1. Fuzzy cognitive map of the pathways of adaptations’ effectiveness or ineffectiveness.

Figure 2. Scenario simulation of the effectiveness of adaptation strategies.

4.5. Contexts of the effectiveness of adaptation strategies

Good weather conditions: Although adaptation decisions are made in response to CCV, the returns of some farm-based adaptation strategies partly depend on optimum rainfall (Figure 1). Farmers may plant early or late following the onset of rain. However, if the amount of rain is not optimal for crop growth, changing planting time may not help farmers to obtain a good amount of yield. Likewise, improved seeds require an optimum amount of rainfall to benefit farmers. They also require the supplemental use of productivity-enhancing inputs such as fertilizer, which is rain-dependent. As farmers noted, “fertilizer is good if the weather condition is good.” [FGD-M-8].

Socio-cultural contexts: Effective adaptation is partly rooted in the socio-cultural contexts of farmers. Effective farmers are characterized by high self-efficacy, risk-taking, and openness to change. Farmers with high self-efficacy, which shows a belief in one’s ability to take adaptation action, were aspired for self-improvement and committed to working hard. A farmer who claimed to be effective explained his personal experience saying that.

“I strongly believe that people cannot change their lives unless they work hard. With this strong conviction, I started working hard, day and night, and my life got improved very well. Thanks to God! I am living in a very good condition.” [IND-L-38]

In addition, effective farmers were not hopeless due to recurrent exposure to CCV. Farmers who were risk-taking and open-to-change, most of whom were better off, were more effective.

Asset contexts: Adaptive capacity, which is the function of access to livelihood assets was, important for effective adaptation (Figure 1). As one key informant noted, “They [farmers] can increase their productivity according to their level of asset ownership.” [KII-H-1]. For instance, farmers who have large size of land benefitted from the use of adaptation strategies as they had options for diversifying crops. Farmers with higher social capital were able to draw resources (e.g. money, labour, improved seeds) to make use of adaptation opportunities and reduce the risks of disposing of key assets. Besides, households who had access to assets properly implemented agricultural extension advice. Access to water was a key physical asset influencing the benefit of adaptation strategies. Farmers in the lowland areas where there is access to Awash River explained that “when we use the full package [of production] on irrigation land, it noticeably increases production. Because there is water.” [FGD-L-11]. Access to water increased the effectiveness of farm-based adaptation strategies by improving the yield-enhancing benefits of fertilizers. “We use fertilizer because there is water on the irrigation farms. Using fertilizer gives you a better yield.” [FGD-L-11]. It also widened opportunities to engage in high-return adaptation strategies such as animal fattening. The utilization of credit services for adaptation investment was also dependent on the availability of water or access to an irrigation canal.

Strategy contexts: The use of multiple and high-return strategies, and proper use of strategies increases effectiveness (Figure 1). Multiple adaptation strategies created opportunities to generate income from different sources. Likewise, engagement in high-return strategies immensely contributed to the improvement of farmers’ livelihoods. Maximum possible benefits were obtained from the strategies when they were used properly. For instance, the benefit of improved seeds was higher when integrated with the use of a sufficient amount of fertilizer. Adherence to agricultural extension advice in the implementation of adaptation strategies was also important for the proper use and effectiveness of the strategies. In particular, the effectiveness of farm-based adaptation strategies such as adjusting crop type or planting time was mainly the function of proper use of farm inputs and extension advice.

Institutional contexts: Formal institutions supported adaptation mainly through the supply of farm inputs (e.g. improved seeds, herbicides, pesticides, and fertilizers). Extension services, particularly farmers’ training, contributed to the effectiveness of adaptation strategies. “It is due to the training and proper planning that I am effective.” [IND-H-1]. Credit service was another means through which institutions supported effective adaptation. One key informant revealed this saying: “I borrowed from a saving and credit association. I worked with it well … So, the credit helped me a lot.” [IND-H-9]. Farmers disclosed that they can respond to climate problems effectively when these institutional supports are adequate and provided at an appropriate time.

4.6. Contexts of the ineffectiveness of adaptation strategies

Climate contexts: Although adaptation strategies are pursued in response to CCV, there is a threshold level of CCV beyond which some strategies are not effective. For instance, CCV affected the benefits of improved seed varieties and fertilizers due to erratic rainfall. Farmers explained that “when there is lack of rainfall, what is spent on fertilizer becomes a loss.” [FGD-M-6]. This quote indicates that the use of fertilizers to increase the productivity of improved seeds is less effective during years of rainfall shortage. In the highland areas, the late onset of rainfall and the consequent delay of sowing time exposed long-duration crops to frost at the ripening stage, resulting in crop failure or substantial yield reduction. Farmers who were highly sensitive to climate-related problems were also discouraged from using credit services. “Our area [lowland] is not favourable to receive credit. In other areas where there is water, farmers can improve themselves by taking credit.” [FGD-L-9].

Fatalism: Adaptation was less effective for farmers with a fatalistic attitude. These farmers were characterized by low self-efficacy and risk-averse behaviour. These farmers attributed their lack of improvement to God’s will and the belief that they cannot go against His will. These households were less inspired to attend farmers’ training, implement knowledge and skill gained through training, and work hard to overcome climate-induced problems. They were also less receptive to new technologies, ideas, and production practices. The barrier to effective adaptation was also related to a lack of trust in the strategies (i.e. strategy scepticism) (Figure 1). Some farmers perceived that they lack the capacity (money, labour, and other resources) to take adaptation action. One informant explained this as follows:

“I have no capacity to cope with the problems of yellow rust and frost. We cannot cover our farm with clothes to protect it from yellow rust. It [frost] destroys vetch and lentil. We cannot protect it. It becomes beyond our capacity.” [IND-H-9]

Risk-averse farmers opted to avoid the use of, or make a little investment in, adaptation with the fear that they would not be successful due to CCV.

Livelihood assets: The adaptation benefit was related to access to key livelihood assets (Figure 2). Given the traditional method of ploughing farmland, farmers who lacked oxen and wait for the support of others to plough their farmland were unable to take advantage of early or late rainfall to plant seeds. Financial problems constrained adaptation investment and effectiveness. Although access to credit is one of the means to access finance required for adaptation investment, the exclusion of the poor who are in dire need of financial resources was a pressing challenge. According to key informants from microfinance institutions (KII-H-5 and KII-M-10), the most widely used procedure to provide credit in rural areas is the group credit scheme. In this scheme, if one member fails to pay the credit, the remaining members are responsible to pay. Hence, better-off farmers often exclude the poor for the fear that they expend money on consumption items and are perceived to be unable to pay the credit. Lack of access to water, due to water scarcity in the highland and midland areas and ineffective governance of water use in the lowland areas, reduced the effectiveness of adaptation strategies by increasing production risks. One farmer from the midland area who was using a small spring for irrigation espoused his experience as follows:

“I use irrigation. About 300 people are using the water for household purposes. Besides, 21 households use the river for irrigation. As the water is not sufficient, we use it in turn. My turn comes every 21st day. By this time, the plant is already dried out. So, it is not beneficial.” [IND-M-12]

Furthermore, small landholdings undermined crop diversification options, and technology adoption varied by the economic status of farmers in which the better-off were the best adopters.

Nature of use of adaptation strategies: The nature of adaptation strategies (e.g. use of low-return strategies, reliance on one strategy, and sub-optimal use) was among the factors shaping the effectiveness of farmers’ response to CCV. In the highland areas, due to the persistent problems of waterlogging and frost, farmers shifted their production decisions from crops such as bean to a crop type locally called shallo, which is resistant to the problem of low productivity. A key informant from the Office of Agriculture (KII-H-1) noted that the crop has very low nutritional value. Farmers relying only on farming activities were less effective. In particular, farm-based strategies were less effective when crops are produced only once a year or only a few crop varieties are produced. Farmers were also less effective when they engage in strategies involving the production of outputs of less market value (e.g. maize, sorghum). Maize cannot be used, as farmers noted, beyond feeding their children for one or two months, and if sold, the money covers only a few household expenses. Sub-optimal use of strategies (e.g. the amount of improved seeds not proportional to the size of farm plots) substantially reduced the benefits to be obtained thereof.

Institutional contexts: The biggest challenge of formal institutions in supporting effective adaptation was related to a lack of, or inadequate and delayed supply of, farm inputs. Consequently, farmers were obliged not to use the strategies, delay the time of taking action, or use inadequate amounts of inputs. Furthermore, the treatments farmers used were not effective in rescuing crops, which, according to them, was related to the poor quality of the inputs. The low capacity of the local formal institutions, governance problems, and poor market functioning contributed to ineffective adaptation. Although three Development Agents (DAs) with three different fields of study (Crop Production, Livestock Production, and Natural Resource Management) are supposed to work in each kebele, the majority of the kebeles in the study area had at most two DAs. The major governance problems, as reported by farmers, include limited support of the DAs, exclusion of the poor from livelihood opportunities and participation in farmers’ organizations, lack of transparency in the distribution of farm inputs, and inequitable access to irrigation canals. Market inefficiency partly explained the limited effectiveness of some adaptation strategies. It mainly shows the imbalance between the price of farm products and the commodities purchased from a market. In the lowland areas where onion is produced and in the highland areas where dairy products are common, farmers noted lower market prices as key constraints of livelihood improvement. Market prices of farmers’ products substantially dropped and farmers failed to cover production costs due to undue interference and influence of brokers and traders. This reduced the benefits farmers could obtain from animal fattening and the production of dairy products and vegetables. There was also an inflated price for purchasing farm inputs which discouraged poor farmers from investing in adaptation. Poor market linkage was another challenge farmers faced to sell their farm and non-farm products.

Non-climate factors: Farmers faced several non-climate risks that influence the effectiveness of their adaptation action. These include conflict and landslide. In one village in the lowland areas, conflict with the neighbouring community limited adaptation investment (e.g. construction of terraces and water reservoir). During the recurrent communal conflicts, humans were killed; animals were looted; and resources were destructed. Consequently, farmers were less confident to invest in their land, leading to precarious livelihood conditions. Landslide was a risk factor in the midland areas. As farmers noted, it resulted in the damage of farmlands, small springs, irrigation canals, terraces, and water reservoirs.

In the cognitive maps, adaptation capacity was found to have the highest centrality, denoting its importance for adaptation effectiveness. This variable (i.e. increasing adaptive capacity) was used to simulate changes in the level of effectiveness of adaptation strategies compared to the steady state of no changes. The steady-state calculation as well as the the values of the simulation were obtained after 20 iterations. Figure 2 shows the results of the scenario simulation. The results show that interventions aimed at increasing the adaptation capacity play significant roles in positively influencing the components of the system. In addition to noticeably improving the adaptation’s effectiveness, it substantially increases self-efficacy, use of technologies, engagement in high-return activities, and optimal adaptation investment. Conversely, increasing adaptive capacity reduces maladaptation, delayed adaptation action, sub-optimal adaptation investment, reliance on few strategies, and engagement in low-return livelihood activities.

5. Discussion

The overriding objective of this study was an in-depth understanding of the contexts of the effectiveness of adaptation. We investigated the factors influencing the extent of effectiveness of adaptation strategies by using a holistic system approach focusing on socio-cultural, economic, institutional, and biophysical contexts. The key findings are briefly discussed below.

The effectiveness of adaptation strategies is embedded in the favourability of rainfall conditions: Although our quantitative impact study shows that adaptation is generally beneficial (Etana et al., 2021a), the effect of CCV is not completely offset by the use of adaptation strategies particularly for poor farmers (Wossen et al., 2018). The results of this study show that rainfall conditions contribute to the ineffectiveness of adaptation strategies both directly and indirectly. If the amount and distribution of rainfall significantly deviate from normal, it may directly cause crop failure or yield reduction. For instance, the false onset of rainfall in Ethiopia challenges the effective use of adjustment of planting dates (Kassie et al., 2013). Rainfall variability indirectly affects the effectiveness of adaptation by decreasing adaptation investments. This result agrees with the findings of other studies in Ethiopia that rainfall variability reduces the use of credit services, fertilizers, and improved seeds (Alem et al., 2010; Kassie et al., 2013; Teklewold et al., 2017). With poor rainfall and the consequent yield reduction, the capacity of smallholder farmers to engage in and benefit from non-farm activities is also weakened.

The effectiveness of some adaptation strategies varies across agro-ecological settings: We found that the effectiveness of adaptation strategies varies across agro-ecological settings. The settings differ in climate problems and agricultural production potentials which necessitate the use of agro-ecologically relevant adaptation strategies (Teklewold et al., 2017). The most effective strategy in the highland and midland areas is livestock production (animal fattening and rearing dairy cows). In the lowland areas, where the length of the growth period is shorter, irrigation is more effective. It contributed to effective adaptation through the creation of opportunities to generate higher income from market-oriented products and production multiple times a year. Terracing and soil and stone bunds support effective adaptation in all areas through the retention of soil moisture. Even if there is a lack of rainfall, the retained moisture reduces crop failure. Improved seeds are moderately effective in all areas. The major problem is that, these seeds require the use of much more fertilizer, more optimal growth conditions, and optimal management compared to what conventional seeds often require.

Fatalism reduces effective adaptation, but intrinsic motivation increases it: The finding shows that farmers with fatalistic attitudes are less effective in responding to CCV. With a fatalistic attitude, farmers make very limited efforts to adapt to CCV which according to Bernard et al. (2011) refers to ‘aspiration failure’. As evidenced in our study, fatalistic farmers are characterized by low self-efficacy, strategy scepticism, risk avoidance, and resistance to change. Farmers with low self-efficacy and sceptical of the effectiveness of strategies make sub-optimal adaptation investments as they believe that their efforts may not protect their livelihoods from CCV. As shown in previous studies, risk-avoiding behaviour, which is common among poor farmers, tends to discourage them from using agricultural extension advice and services (Alemayehu et al., 2018; Kassie et al., 2013). Instead, they depend on traditional livelihood practices and involve in low-return livelihood strategies. If adaptive capacity is perceived to be low, farmers’ motivation for and willingness to take adaptation action is reduced. On the contrary, intrinsically motivated farmers are highly likely to respond to CCV effectively. Motivational factors mainly include high perceived self-efficacy, openness to change, and risk-taking. As shown in previous studies, these factors are important for effective adaptation (Etana et al., 2020a; Gebrehiwot & van der Veen, 2021). Farmers with these attributes believe they can change their livelihoods using all possible means including adaptation. They strive to overcome poverty instead of accepting and remaining to live in their current living conditions.

Adaptation is not equally effective for all households due to differential adaptive capacity: Our finding shows variation in the effectiveness of adaptation strategies by farmers’ adaptive capacity. This finding augments the results of previous studies underscoring the important roles of access to livelihood assets for adaptation investment (Bryan et al., 2009; Di Falco et al., 2011; Etana et al., 2020a; Gebrehiwot & van der Veen, 2013). The adaptation decisions of poor farmers lack flexibility and their decisions fail to meaningfully improve their livelihoods. Farmers with limited access to assets resort to low-cost and low-return adaptation strategies which are less impactful in improving livelihoods. We also found that low adaptive capacity increases maladaptation. The adaptation decisions of farmers with a low adaptive capacity, for instance, deforestation for charcoal production and children’s school drop-out, increase cyclical vulnerability and a vicious circle of poverty. On the other hand, farmers with a high adaptive capacity respond to CCV effectively through flexible use of technologies and engagement in high-return adaptation strategies. The results support the evidence that increasing adaptive capacity is at the core of effective adaptation (Bryan et al., 2009; Teklewold et al., 2017; Wossen et al., 2018). The effectiveness of adaptation among the better-off farmers implies a shift of vulnerability to poor farmers, consequently increasing inequality in the community.

Diversification of adaptation strategies increases effectiveness: Farmers who use multiple adaptation strategies and engage in high-return strategies respond to CCV effectively. The finding confirms the results of our quantitative impact study (Etana et al., 2021a) as well as other studies in Ethiopia (Di Falco & Chavas, 2009; Teklewold et al., 2017) showing the benefits of using multiple adaptation strategies. Diversification of sources of income increases household welfare and decreases the risk of falling into poverty (Kidane & Zegeye, 2020), and reduces the impact of climate and price variability (Wossen et al., 2018). The benefits of using multiple adaptation strategies are associated with the synergy and complementarity between different strategies (Teklewold et al., 2017).

Market functioning determines adaptation effectiveness: The results of the study show that a poorly functioning market reduces the effectiveness of adaptation strategies which is consistent with earlier works identifying price volatility as a key challenge (Kassie et al., 2013). Farmers in the study area are strongly integrated into the market which makes their livelihoods vulnerable to market-related problems. They increasingly participate in market-oriented adaptation strategies such as animal fattening, production of vegetables, and supply of dairy products. However, the price of these products is unfairly influenced by brokers. Consequently, farmers fail to get a reasonable benefit from using the strategies. Besides, due to fluctuation in crop yields induced by rainfall variability, farmers rely on grain purchased from the market. However, commodities are bought at a high price whereas farm products are sold at a lower price (Leta et al., 2017). Farmers who depend on low-return strategies are therefore affected by this price imbalance as they earn less but spend more. Other studies similarly acknowledged that poor farmers who spend a higher proportion of their income on the purchase of food items are highly vulnerable to the increasing prices of consumption goods (Sabates-Wheeler & Devereux, 2010; Wossen et al., 2018) and face severe food shortage (Matz et al., 2015). Farmers’ producer organizations are the major means through which farmers can overcome market problems. In our study, however, poor farmers are less involved due to socio-economic marginalization.

The efficiency of institutional support of adaptation and governance determine the effectiveness of strategies: The findings of the study show that the role of local formal institutions for effective adaptation depends on efficient institutional support and good governance. Agricultural extension service (e.g. farmers’ training) is one of the means through which institutions support effective adaptation. This confirms the results of previous studies stating that extension services are the key constituents of effective adaptation (Bryan et al., 2009; Gebrehiwot & van der Veen, 2013). One of the key institutional problems undermining the effectiveness of adaptation strategies is the unmet need for farm inputs in terms of timely and adequate supply, equitable distribution, and quality of inputs. The result signifies the importance of the timely supply of inputs for effective adaptation (Arslan et al., 2015). When farmers are unable to meet their input needs, they are obliged not to use the inputs, buy from traders at a very high price, purchase poor quality inputs with cheaper prices, or make a sub-optimal investment, all contributing to the ineffectiveness of adaptation. We found that the paramount roles of the DAs in supporting effective adaptation are constrained by poor-quality extension services. A recent study similarly revealed the lack of motivation and technical competence of the DAs and the inadequate facilities for demonstration at the farmers’ training centres as the challenges of the agricultural extension system in Ethiopia (Hailu et al., 2020). Our finding further shows poor governance as a cause for ineffective adaptation. It is manifested in the marginalization of the poor from adaptation supports and inequality in access to irrigation infrastructure. Consequently, institutional services are mostly utilized by the better-off farmers, which create inequality in adaptation effectiveness.

Strengths and limitations: The main strength of this study is that the analysis is based on rich textual data that gave depth to the explanation of the reasons adaptation strategies are effective or not. Quantitative studies show the statistical relationship between the use of adaptation strategies and livelihood outcomes. However, this qualitative study helped to further explore the micro details on the pathways of effective and ineffective adaptation decision-making which are less amenable to precise measurement using a quantitative approach. The limitation of the study is that the relationship between the variables was determined based on normative judgements. However, the effect of this on the validity of the results is minimal as the findings of the FCM are supported by in-depth understanding obtained from the thematic analysis of FGD, IND, and KII data as well as several other empirical studies.

6. Conclusion

This study investigated the contexts that affect the livelihood benefits of adaptation strategies. The following key conclusions are drawn from the findings. First, adaptation is not independent of the effects of CCV. In particular, the effectiveness of rain-dependent adaptation strategies is conditional on the optimum amount and distribution of rainfall. Second, the adaptation decisions of poor farmers lack the flexibility to take advantage of existing adaptation opportunities. On the other hand, farmers with a very strong work ethic founded on intrinsic motivation are more likely to respond to CCV effectively. Third, adaptation strategies may help farmers only to reduce vulnerability but may not help to overcome the livelihood threats posed by CCV. For instance, the uses of non-farm and migration as adaptation strategies help in covering subsistence costs but not in livelihood improvement. Fourth, the beneficial impact of using adaptation strategies varies by farmers’ adaptive capacity. For better-off farmers, adaptation is helpful to build sustainable livelihoods but for poor households, it may not noticeably reduce vulnerability or in the worst case increase vulnerability and inequality. Building sustainable livelihoods through adaptation is far from being realistic for farmers with low adaptive capacity. Fifth, efficient institutional support schemes and well-functioning markets are important for effective adaptation.

The findings have implications for policy and practice to support effective adaptation. It is important to identify contextually relevant and agro-ecologically feasible adaptation strategies to maximize their benefits. For instance, the expansion of irrigation infrastructure in the lowland areas where it can be sustainably used is helpful to enhance sustainable livelihoods. In the midland and highland areas, supplying improved animal breeds and building farmers’ skills in modern production practices are important to increase the benefits of livestock production. It is worth noting that the environmental impact of livestock production, as well as the downstream impact of irrigation, shall be taken into account to avoid maladaptation. The dependence of farm-based strategies on optimum rainfall necessitates the dissemination of information on the time of onset and cessation of rainfall to help farmers make informed adaptation decisions. Building the capacity of farmers, for instance, addressing the credit constraints of poor farmers is one of the means to ensure effective adaptation to CCV. This increases overall community resilience by narrowing the gap between effectively adapting better-off farmers and ineffectively adapting poor farmers. Institutions can support effective adaptation by establishing an efficient system for a timely and adequate supply of farm inputs. Extension support on and creation of market linkage for livestock fattening and other market-oriented strategies in all areas; establishing a well-functioning value chain of supply of agricultural products; and increasing the use of technologies that can efficiently link the producers and consumers immensely help in improving the benefits farmers obtain from the use of adaptation strategies.

Acknowledgements

The authors are very much grateful to the research participants for their willingness to share their knowledge and experience. We also thank the anonymous reviewers and the editor for their critical comments.

Disclosure statement

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

Additional information

Funding

This work was supported by EP-Nuffic [grant number: R/002597.01].