4.1. Choice of model
For a farmer to make choice on whether or not to adapt to a particular technology or innovation, he does not only consider how to maximize profit from that innovation but on how to attain the highest level of utility otherwise referred to as utility maximization (Grabowski et al. 2019). It is observed that farmers have a level of utility they want to meet and therefore make choices based on that. Any adaptation option could fall under the general framework of utility and profit maximization (Kassie et al. 2008). The utility of a farmer is given as Uij, from choosing alternative j. Farmers will choose whether or not to adapt to climate change depending on the relative utility levels associated with the two choices. In this case, choice models are used to analyze a farmer’s decision to adapt or not to adapt to climate change. This model assumes that farmers are likely to adopt a novel strategy only when an economic benefit obtained from such a strategy is significantly greater than from their actual strategy (Khan & Akhtar 2015). Although farmers’ profit maximizing behaviour using adaptation strategy is unobservable directly, their actions can be understood indirectly through choices they make (Bedeke et al. 2019).
This study employed a Multinomial Logit (MNL) model. This model is advantageous over other Multinomial Probit (MNP) model in that it allows predicting adaptation choice in terms of odds ratio with probabilities. The MNL is widely employed in climate change adaptation studies (Hisali et al. 2011; Tessema et al. 2013; Gbetibouo 2009; Gebrehiwot and Van Der Veen 2013). The MNL technique compares any given outcome with a reference outcome. This technique is deemed suitable to study adaptation to climate change since households employ different adaptation strategies which are typically not mutually exclusive. It is employed when the dependent variable has more than two outcomes or, in our case, more than one adaptation response to climate change (Greene 2012). To describe the MNL model, let y denote a random variable taking on the values {0, 1,…..,M} where M is a positive integer and x denote a set of conditioning variables. In our study, y would be adaptation measures taken by households whereas x represents the explanatory variables hypothesized to influence the choice of the available adaptation options. The MNL model is employed to show how ceteris paribus changes in the elements of x influence the response probabilities, P (y = m/x), m = 0, 1,…, M. P (y = m/x) is known after determining the probabilities for m = 0,1,2,…,M, which must sum to unity. The MNL model has response probabilities given as Equation (1):
![](data:image/png;base64,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)
The parameter estimates of the MNL model only show the direction of the relationship between the dependent and independent variables. Therefore, to determine the actual magnitude of change or probabilities, the marginal effect of the explanatory variables, Equation 1 is differentiated over the explanatory variables to give Equation (2):
![](data:image/png;base64,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)
The MNL, however, works under the assumption of the Independent Irrelevant Alternatives (IIA). Following this assumption, the odds of any two outcomes are independent of the remaining outcomes available. Hence, omitting or adding outcomes should not affect the odds of the remaining outcomes (Hisali et al. 2011). The fitted MNL model was first checked to make sure that it does not violate this assumption.
4.2. Model variables
The fact that climate has changed in the past and will continue to change in the future underlies the need to understand adaptation processes, including its determinants and drivers (Hisali et al. 2011). Assessing farmers’ climate change adaptation choice is important as several climate studies ask individuals about their choice to new practices that build resilience. Farmers’ adaptation choices are driven by diverse institutional, socio-economic, behavioural and factors (Grothmann and Patt 2005; Woods et al. 2017).
Dependent variable
The dependent variable in this study is whether a household has ‘adapted’ or ‘not adapted’ any adaptation practices to climate change. Based on discussions with development staff, review of the literature and field observations, the adaptation practices identified included improved crop varieties (drought-tolerant and early maturing crops), crop diversification (mixed cropping and crop rotation), farm diversification (mulching, composting, ridging and terracing), change in planting date, income generating activities, irrigation practice (dry season gardening) and agroforestry. Adaptation choice is the dependent dummy variable. To determine the dummy, a value of ‘1’ was assigned to those households that had adopted at least one of the adaptation options and ‘0’ for those that had not adopted.
Explanatory variables
Explanatory variables hypothesized to influence the choice of adaptation strategies by crop producers are provided in detail.
Institutional factors
Institutional factors can include household access to weather information, agricultural input price information and access to agricultural training service. These factors have been widely used as variables in several adoption studies in order to evaluate farmers’ behaviour (Bryan et al. 2009; Deressa et al. 2009; Mudzonga 2012). Farmers’ access to information on weather change may have a significant relationship with their choice of climate change adaptation strategies (Nguezet et al. 2013). Exposure to extension services and mass media increase the awareness and concern regarding climate change. Such information obtained through extension agents or mass media (radio or television) can help promote crop and varietal choices (Grabowski et al. 2019). Access to detailed information on local market prices of improved crop varieties and pesticides is hypothesized to influence farmers’ livelihood diversification choice (Nhemachena et al. 2014). This is because access to adequate market prices of agricultural inputs may be associated with their capability to undertake decisions to diversify into non-farm income strategies (Mulenga et al. 2016). Access to agricultural training services is found to influence farmers’ adaptation decisions by enhancing their capacity to acquire new experience in dealing with climate change (Piya et al. 2013). Access to microcredit services may ease farmers’ cash constraints and thus positively associated with choice of adaptation strategies such as chemical fertilizers, high-yielding crop varieties, and/or irrigation pumps (Nhemachena et al., 2014; Wainaina et al., 2016).
Socio-economic factors
Several studies used socio-economic factors such as education levels, farm experiences, land size and gender to explain household adaptation choices (Asfaw et al. 2013; Shiferaw et al. 2014; Boansi et al. 2017). A household education level is associated with access to output market price information as well as increased income that influence farmers’ adaptation choices. Most climate change studies tend to agree that educated households are more able to process agricultural inputs, allocate them efficient, assess the profitability of the new strategy by taking into account costs and subsequently undertake decisions to adopt adaptation than illiterate households (Nguezet et al. 2013; Grabowski et al. 2019). With reference to the farm experience which is mainly related to age, several studies are of the view that households with high levels of farm experience are likely to undertake adaptation decisions (Bryan et al. 2009; Deressa et al. 2009; Falco 2014). Large land size may enable households to undertake a variety of agricultural activities and hence is likely to affect farmers’ adaptation decision (Kassie et al. 2013). Gender of the household may influence potential adoption of adaptation strategies to climate change (Ndiritu et al. 2014). Household non-farm income can be positively correlated with climate change adaptation choices. Increased household non-farm income from petty trading, woodworking and animal bartering provides farmers with additional financial capital for investing in improved crop varieties and fertilizers (M. Kassie, Teklewold, and Jaleta 2015). However, depending on household socio-economic and demographic contexts, climate change studies that examine the effect of gender differ on whether female or male household heads are more likely to adopt adaptation strategies (Asfaw et al. 2013; Falco 2014).
Behavioural factors
Several studies underscore the importance of farmers’ behavioural factors such as risk perception in explaining adaptation choices (Adimassu et al. 2014; Niles and Mueller 2016; Sutcliffe et al. 2016). Farmers who perceived climate change risks in terms of increased temperatures are more likely to believe negative atmospheric warming that influence their choice to use heat and disease tolerant crop varieties (Niles and Brown 2016). Perceptions of declined annual rainfall risk is hypothesized to have a positive and significant relationship with choice of climate change adaptation strategies (Mertz et al. 2009). More frequent droughts and lack of sufficient rainfall risks driven by climate change are expected to adversely affect farmers’ productivity and yields, and could subsequently influence their choice of adaptation (Kotir 2011; Shiferaw et al. 2014). Farmers’ perceptions of increased seasonal rainfall variability patterns may have positive relationship with their choice of adaptation to climate change (Nguezet et al. 2013; Grabowski et al. 2019). Farmers who concerned with more increase in overall negative climate effects on their crop yields are likely to undertake adaptation choices (Niles and Brown 2016).
Socio-cognitive factors
The social-cognition approach assesses several forms of people’s knowledge about global climate change that they have acquired through their long-term experience. Local people’s beliefs in terms of climate change influence their adaptation choices in their farming and livelihood practices. These beliefs can be explained by local people’s awareness of whether climate change is occurring and caused by human actions and/or natural events (Hyland et al. 2015). In northern Ethiopia, Tesfahunegn et al. (2016) showed that farmers’ tree planting decisions in highly degraded lands may be influenced by their beliefs that climate change is caused by deforestation. As postulated by the theory of social-cognition that deals about human agency (Bandura 2001), local people’s beliefs in climate change form the basis for critical information that bring changes in their attitudes following perturbation, and ultimately leads to explain adaptation choices.
However, increased climate uncertainty may lead to either absolute belief in or total denial of climate change (Abegunde 2017). Farmers can be either “believers” who have knowledge about its causes and occurrences or “disbelievers” who deny its existence. There are also some farmers who may choose to be neutral in their position towards climate change occurrence (Helgeson et al. 2012). Abegunde (2017) identified three sets of disbelievers in climate change cause, occurrence and effect: trend, attribution and impact sceptics. Trend sceptics are those people who never believe in climate change occurrence. Attribution sceptics accept the occurrence of climate change, but do not believe that this change is caused by a human action. Impact sceptics believe in climate change cause and its occurrence but deny its damage. For example, households may believe in the cause and occurrence of climate change, but they may find evidence of its impacts occurring elsewhere (Grothmann and Patt 2005). To this end, in this study, farmers’ beliefs in climate change occurrence and risks is hypothesized to affect adaptation choices both positively and significantly (Helgeson et al. 2012).
Table 1
Mean and standard deviations of dependent and explanatory variables used in the multinomial logistic model (n=340).
Variable name
|
Variable description
|
Mean
|
SD
|
Access to climate information
|
= 1 if a farmer has access to climate information, 0 otherwise
|
0.58
|
0.37
|
Access to input price information
|
= 1 if a farmer has input price information, 0 otherwise
|
0.72
|
0.52
|
Access to micro credits
|
= 1 if a farmer has microcredit access, 0 otherwise
|
0.89
|
0.59
|
Access to training on adaptation
|
=1 if a farmer receives training on agriculture, 0 otherwise
|
0.48
|
0.36
|
Belief on climate change
|
=1 if a farmer concern over climate change, 0 otherwise
|
0.73
|
0.45
|
Crop and varietal selection
|
=1 if a farmer changed crop and variety use, 0 otherwise
|
0.82
|
0.53
|
Education level (in years)
|
Number of years of schooling
|
2.30
|
1.38
|
Farm experience (in years)
|
Number of years of experience in farming
|
35.0
|
2.82
|
Farmers’ perception of increased drought risks
|
= 1 if a farmer perceives increased drought risks, 0 otherwise
|
0.84
|
0.29
|
Farmers’ perception of increased rainfall intensity
|
= 1 if a farmers perceives increased rainfall intensity, 0
otherwise
|
0.58
|
0.39
|
Farmers’ perceptions of increased erratic rainfall
|
= 1 if a farmer perceives increased erratic rainfall, 0 otherwise
|
0.83
|
0.40
|
Gender of the household head
|
=1 if a household is male-headed, 0 otherwise
|
0.79
|
0.68
|
Access to extension contact
|
=1 if a farmer is visited by extension per week, 0, otherwise
|
0.67
|
0.23
|
Household non-farm income
|
Average household non-farm income
|
3234
|
234
|
Household on-farm income
|
Average on-farm income
|
9420
|
340
|
Livelihood diversification
|
=1 if a farmer diversify livelihood options, 0 otherwise
|
0.87
|
0.45
|
Shift farming practice
|
=1 if a farmer shift farming practice, 0 otherwise
|
0.78
|
0.39
|
Size of farm (ha)
|
Average land size
|
0.42
|
0.23
|
Note that data were obtained from household survey 2018 to calculate the overall mean and Standard Deviation (SD) for the dependent and explanatory variables. Some explanatory variables were made dummy (1=yes, 0=otherwise) for the sake of easier comparison in the MNL model though they are asked in terms of three-point scale. |