Introduction

In this article, we look into the extent to which gender is considered in the climate adaptation policy debate and explore how and why a particular way of framing becomes prominent. In so doing, we assume that variation in the extent to which gender is given a role in climate adaptation policy can be explained by the interactions among policy actors in constructing discourses and, subsequently, policy.

Climate change has become a prominent agenda item both in academia and policy making arenas; consequently, the development of policies and strategies vis-a-vis adaptation has grown (Ford et al. 2015; Lwasa 2014; Massey and Huitema 2016). Recently, the nexus of gender and climate change (adaptation) has become a subfield that is receiving a growing amount of attention in the literature highlighting that climate change problems and, subsequently, responses are not gender neutral. Climate change has differentiated effects on actors that are influenced by asymmetric relations. Asymmetric relations based on social, economic and political grounds can be related with age, education, class, ethnicity, etc. Here, we focus on gender relations as the root for a particularly persistent asymmetric relation that influence men’s and women’s respective adaptive capacities and adaptation choices (Pearse 2016; Skinner 2011; Terry 2009). Thus, it has been suggested that those developing and implementing adaptation policies and strategies need to consider these asymmetric gender relations and their effects (Bee et al. 2013; MacGregor 2010; Skinner 2011). The failure to do so, according to MacGregor (2010), would make any response to climate change insufficient, unjust and, therefore, unsustainable.

However, an overview of recent studies shows that gender dimensions are still not fully considered in climate adaptation policy responses. In their analysis of 31 Sub-Saharan African countries’ National Adaptation Programmes of Action (NAPA), Holvoet and Inberg (2014) found limited gender sensitivity. Magnusdottir and Kronsell (2015), in their study of climate policy in Scandinavian countries, also noted that despite an overall gender balance among policy actors, gender did not get sufficient attention in the shaping of policy responses. Even when adaptation policies do attempt to deal with gender, they tend to be limited to the addition of gender-sensitive language to existing policies without capturing and addressing the underlying factors (Bee et al. 2013; Skinner 2011). Based on such observations, studies warn that adaptation policies may perpetuate or even reinforce gender inequality (Bee et al. 2013; Arora-Jonsson 2011; Pearse 2016).

Going beyond pointing out the differentiated effects of climate problems, gender and feminist scholars also challenge the discursive framing of climate policies that conceals the gender dimension (Israel and Sachs 2013; MacGregor 2010, 2014). Literature indicates that the domination of masculine and apolitical construction of climate change discourses not only underplays gender but also creates a dominant techno-scientific definition of climate problems as best managed through technological and engineering innovation (ibid). The role of discourses in policy processes has been amply studied and discussed (Fischer and Forester 1993; Hajer and Wagenaar 2003). However, despite a relative recent advance in climate studies, there is still a dearth of feminist involvement in the debate about the way in which gender is contemplated in discourses that lead to the creation of institutions and to the formulation and implementation of climate adaptation policies (MacGregor 2014). As a result, the urgent need for a gender analysis that captures ‘the discursive and cultural constructions of hegemonic masculinities and femininities that shape the way we interpret, debate, articulate and respond to’ climate change is suggested (MacGregor 2010, p 127). Therefore, this article aims to contribute to the literature by analysing the interactive discourses among policy actors and the influence thereof on the extent to which adaptation policies addresses the gender dimension.

According to Schmidt (2008), (policy) discourse consists of both the ideas produced (substantive content) and the interactive processes through which these ideas are constructed and conveyed. In discursive policy processes, a variety of sometimes complementing but often competing ideas partake, but a prominent discourse often emerges and dominates the process, making particular things visible while others remain unseen (Bäckstrand and Lövbrand 2007; Schmidt 2008, 2010). Understanding why and how certain ideas—about, for example, climate change, vulnerability and adaptation—become prominent in a given discursive policy process requires examining the process that produces or constructs such ideas and the interactive processes that lead to their rise to prominence (Schmidt 2008).

The article focuses on national-level climate policies. Most studies on the adaptation and gender debate have provided local-level gender analyses of vulnerability and adaptive capacities, which is most important since adaptation is indicated as inherently a local phenomenon (Hilde 2012). But attention should also be given to national-level engagements since adaptation policies and strategies framed at this level steer the local practices (cf. Ireland and McKinnon 2013). Ethiopia is the empirical base of this study. In 2015, the Notre Dame Global Adaptation Index (ND-GAIN) indicated that out of 185 countries, Ethiopia was among the top 40 countries that are most vulnerable and least ready to respond to climate change (ND-GAIN 2015). The recent engagement of the Ethiopian government in both the global and the African climate change debate is remarkable. At the national level as well, Ethiopia has issued a number of climate change adaptation policies and strategies, including a ‘Climate-Resilient Green Economy Strategy’ that aims to make Ethiopia a ‘green economy frontrunner’ (FDRE 2011; Jones and Carabine 2013). This study deals with Ethiopia’s Climate Resilience Strategy in Agriculture and Forest.

Theoretical Foundation

Discursive institutionalism

Discursive institutionalism (DI) is an approach that combines insights from new institutionalism (NI) and discourse theory. According to DI, discourses encompass the policy concepts and ideas and the interactive processes that produce policy (Schmidt 2008, p 305). In DI, ideas, discourses and institutions are considered as the core components.

Ideas are policy contents that carry discourses (den Besten et al. 2014). They are found at three nested levels: (i) in specific policies or policy solutions that are often found at the surface level of policy documents; (ii) in more general ‘programmes’ that serve as a frame of reference through which policy solutions are linked with policy problems; and (iii) in the philosophies or deeply rooted world-views, values and principles that uphold both policy problems and solutions (ibid; Schmidt 2008, p 306).

Interactive discourse refers to the process through which policy actors construct and exchange their ideas, often by means of contestation (Schmidt 2008, p 305). The analysis of the interactive discourse asks how specific policy ideas are conveyed, to whom, when, where and why (ibid). Schmidt (2008, 2010) proposes a differentiation between coordinative discourse (the interactions of policy actors during the policy making process) and communicative discourse (the interaction between policy actors and the public). In our study, we focus on the coordinative discourse, as the communicative discourse arena seems politicised and constrained in Ethiopia.Footnote 1

Institutions entail both formal and informal rules of the game—norms, laws, regulations and procedures—that guide interactions between actors and subsequently how ideas are conveyed, to whom, when, where and why. Institutions are perceived as both ‘given’ and ‘contingent’ (Schmidt 2008, p 314). In other words, institutions create an enabling or constraining environment for the legitimisation of policy ideas, but through their interaction with discourses, institutions are subject to change.

Highlighting the gendered nature of both discourses and institutions, feminist scholars argue for a critical evaluation of how environmental problem discourses are constructed and contested and how institutions reinforce or challenge inequality from a gender perspective (Arora-Jonsson 2011; MacGregor 2010, 2014). For instance, MacGregor (2010) notes that men’s predominant control of climate science results in a discourse that is heavily dominated by masculine values. In institutional studies, the concept of ‘gendered institution’ is used to refer to the gendered nature of the institutional establishment, functioning and change (Lowndes and Roberts 2013; Mackay et al. 2010; Waylen 2014). Viewing institutions as being gendered implies that institutions maintain and reinforce existing gender asymmetric relations (Mackay et al. 2010). Also, when institutions change, unless critically examined and explicitly considered, this asymmetric relationship is sustained (Waylen 2014), although it then often remains hidden under a cover of ‘gender neutral’ language. One aspect of the debate on the subject focuses on how institutions change and how this change could lead to the emergence of ideas, interactions and (new) institutions that could bring about more gender-responsible outcomes (ibid). In this study, we opt for DI as a promising approach to understanding the role of gender in policies (i.e. contents) and policy processes (cf. Kulawik 2009).

Discursive institutionalism in the research

Gender, here, refers to a relational and discursive framing that influences social life (MacGregor 2010). This conceptualization shifts gender analysis from focusing on categorical attributes to a broader structure and cultural interpretation (Scott 1986). The premise of DI is that policy making is a process where specific policy ideas gain prominence over others. Hence, studying the interactive discourse helps to understand why some ideas succeed while others fail (Schmidt 2008, 2010). Inspired by DI, the study develops a twofold objective: first, identifying the extent to which gender is represented in adaptation strategies and, second, exploring the way in and the extent to which a gendered coordinative discourse affects policy outcomes.

Regarding the first objective, DI’s three nested levels of ideas (policies, programmes and philosophies) are used to explore and establish the representation of gender in Ethiopia’s adaptation strategies (see Fig. 1). The three nested ideas are dealt with by creating a link to three key questions, raised by Smit et al. (2000), in adaptation analysis, namely (i) adaptation to what and why; (ii) who or what adapts; and (iii) how does adaptation occur. These questions help to evaluate adaptation merits and understand values connected with adaptation (Thomsen et al. 2012). However, gender may not appear at the surface in such lines of questioning because, as Kronsell (2006) noted, a gender-neutral approach often conceals the hegemonic masculine values and biases. Consequently, unpacking gender requires the deconstruction of narratives that normalise the silence by looking into what is not said or written or ‘silenced’ also (ibid, p 109).

Fig. 1
figure 1

Gender and the discursive institutionalism framework

Full size image

Philosophical ideas refer to ‘adaptation to what and why?’ questions. These ideas focus on the overarching perception of climate change, on its gender dimensions and on the specific aspects of climate change problems and other non-climate problems that adaptation is proposed to respond to. Programmatic ideas are related to the framing of policy problems and refer to the question of ‘who or what adapts?’ The intention is to establish what or who is vulnerable and needs to adapt, and as such to determine if or to what extent the gender dimension is captured in problem definition. Policy ideas frame policy solutions and deal with how adaptation occurs. This level focuses on the type and nature of adaptation options and the extent to which they consider gender.

The second objective relates to the analysis of coordinative discourses. As Schmidt (2008, 2010) states, the key questions in this analysis relate to how policy ideas are conveyed, when, to whom and why. In the quest to answer these questions, we particularly focus on policy actors and their interaction (see Fig. 1). Gendered policy actors deal with the question of ‘to whom’ ideas are conveyed. In policy processes, the dominance of certain ideas does not always come from the powerful messages they carry, but rather from powerful actors who choose to make use of them (Wodak and Meyer 2001). Hence, identifying actors who are involved in the process, as well as their respective agendas and interests in relation to gender, becomes important in order to understand the coordinative discourse.

Gendered interactions deal with the questions of ‘how’ gender policy ideas are conveyed, ‘when’ and ‘why’. In their prominent work, West and Zimmerman (1987) assert that gender is produced and reproduced in every interaction. Hence, to capture the gender aspect of the coordinative discourse, it is imperative to analyse the policy interactions as a gendered process (see also Kronsell 2006). Gendered institutions refer to both formal and informal rules of the game. For instance, policies, programmes, guidelines and informal organisational norms and networks play important roles in facilitating or constraining interaction (Schmidt 2008, 2010); hence, the focus is to identify both formal and informal institutional settings where the inclusion of gender is either promoted or neglected.

Methodology

We apply our discursive institutionalist framework to the case of Ethiopia’s Climate Resilient Green Economy (CRGE) Vision and, more particularly, to the Climate Resilient Strategy for Agriculture and Forest (CRS-AF). CRGE is focused upon because it is a comprehensive national strategy regarding Ethiopia’s response to climate change and the first ever document that has situated climate change within the national development goals (Eshetu et al. 2014). CRGE has three objectives: (i) fostering economic development and growth, (ii) ensuring abatement and avoidance of future GHG emissions and (iii) improving resilience to climate change (FDRE 2011). Ethiopia’s CRGE consists of mitigation and adaptation strategies, notably the Green Economy Strategy (GES) and the Climate Resilient Strategy (CRS). The national GES was issued along with the CRGE Vision in 2011, and so far, two sectoral CRSs have been issued, namely CRS-AF and CRS in Water and Energy in 2015. CRS-AF is selected as the focus of our study because it is comprehensive and incorporates all previously issued adaptation strategies in the agriculture sector (FDRE 2015).

In an attempt to explore the coordinative discourse, we conducted 20 semi-structured interviews with policy actors in government ministries and staff members of local civic society groups and international organisations. Twelve semi-structured interviews were held with government policy actors. Five were from the Ministry of Environment, Forest and Climate Change (MoEFCC) and seven form the Ministry of Agriculture and Natural Resources (MoANR) including experts of Women’s Affairs Directorates found in both ministries. The remaining eight semi-structured interviews were conducted with experts of local research institutes and civic society organisations that have demonstrated an interest in climate policy ‘…such as Forum for EnvironmentFootnote 2 (two interviewees), Lem Ethiopia,Footnote 3 the Forum for Social Studies,Footnote 4 the Institute for Sustainable Development,Footnote 5 the Global Green Growth Institute,Footnote 6 CARE-EthiopiaFootnote 7 and Oxfam GB.Footnote 8

We selected interviewees using purposive and snowball sampling. In the first instance, sampling was purposive, whereby key government actors who are known to be owners of the climate agenda within the MoEFCC and the MoANR were selected. The second stage involved a snowball sampling technique. In this procedure, key informants from each of the government actors were asked to recommend individuals who actively engaged in policy making.

We used qualitative content analysis for exploring the data gathered from both document reviews and individual interviews. The analysis of written policy texts and interview transcripts followed two steps: coding and thematic grouping. Coding refers to ‘attaching labels to segments of data that depict what each segment is about’ (Charmaz 2006, p 3). Verbatim records were read to identify key statements and phrases on climate change causes, impacts and responses and the gender nexus. These were coded in relation to the three levels of ideas (policy, programme and philosophical ideas) using both manifest and latent coding. According to Graneheim and Lundman (2004), manifest coding focuses on the surface and visible context of the text, whereas latent coding focuses on the underlying meaning of the text. Then, we engaged in thematic grouping, which is the systematic organisation of contents and ‘a way to link the underlying meanings together in categories’ (ibid, p 107). Direct references to organisations or interviewees in the text are avoided; instead, issues are discussed confidentially in order to ensure the anonymity of the respondents.

Research context: gender in Ethiopia

With a population of 90.1 million (CSA 2015), Ethiopia is socially diverse, comprising more than 80 ethnic groups. This diversity necessitates understanding men’s and women’s different experiences across ethnic groups, livelihoods, localities and classes. Nevertheless, by and large, masculine valuesFootnote 9 are favoured, while women and their experiences are subjugated in most parts of the country (Levine 2000). Gender remains an important factor in defining entitlements to resources, power relation and access to institutions, information and decision-making (UNDP 2015).

Since the regime shift in 1991, the formal institutions targeting gender equality and women’s rights have evolved from attaining legal status to playing a proactive role in voicing equality. In particular, the Constitution of the Federal Democratic Ethiopia (FDRE 1995) recognises the equality of women as equal to those of men in political, economic and social spheres. Article 35 (6) of this Constitution, in particular, stipulates that:

[w]omen have the right to full consultation in the formulation of national development policies, the designing and execution of projects, and particularly in the case of projects affecting the interests of women.

Apart from this, progressive measures have been taken to consider women’s rights in other laws, legislations and policies such as the Rural Land Administration and Use Proclamation (1997, revised in 2005), the Federal Family Law (2000) and the national development plan, namely GTP I and II. Moreover, Women’s Affairs offices have been put in place in most ministries to facilitate the process of gender mainstreaming in government policies, programmes and practices (Biseswar 2008; Østebø 2015).Footnote 10

Given these institutional and legislative grounds, one might think that pronouncing a ‘gender equality’ agenda in any public policy arena including climate change is a default venture. However, as widely acknowledged by official reports, an implementation gap impedes progress within formal institutions (ibid), while informal institutions remain highly gendered (Mersha and Van Laerhoven 2016).

Analysis and results

Substantive content of the discourse: gender in climate policy (ideas)

When evaluated from a gender perspective, the first observation is that the CRS-AF and CRGE texts rarely mention ‘gender’ and provide no elaboration on how the gendered nature of climate change problems and the solutions thereof can be addressed. While an assessment of gender responsiveness needs to go beyond counting frequency of words related to ‘gender’, ‘women’ or ‘men’, this can nevertheless give an indication of potential shortcomings of a policy. In the CRS-AF document, ‘gender’ is referred to four times, but when it is, it remains unclear what is actually meant. The same is true for the word ‘women’. The document refers to ‘women’ twice, portraying them as vulnerable groups eligible for a specific adaptation option, namely social protection. Besides these mentions, the documents treat men and women as homogenous groups and use a ‘gender neutral’ language.

The discussion below provides a more detailed examination of the document regarding the three types of ideas.

Philosophical ideas: adaptation to what and why?

Regarding the question of adaptation to what, CRS-AF provides a comprehensive and broader description of climate-related triggers for which adaptation responses are carried out. Climate stimuli consist of changes in temperature and rainfall, climate variability and extreme weather events such as drought and flooding and to present as well as future contexts are discussed. By and large, adaptation is framed as a response to exogenous climate-change problems in the CRS-AF document despite some mention of the role of non-climate drivers as indicated in the following quote.

Other non-climatic drivers are inextricably linked to vulnerability but have not been explicitly included here in order to distinguish between largely ‘exogenous’ drivers (determined by climate change), and largely ‘endogenous’ drivers, which economic policy can strongly influence (FDRE 2015, p 26, italics added).

The emphasis given to ‘exogenous’ climate change drivers obscures gender from part of the discussion about problem definitions because such framing underplays the non-climate socio-economic, political and institutional drivers where the gender asymmetric relations have entrenched upon and emanated from. Jones and Carabine (2013) also observe the minimisation of complex sociocultural and political factors in the ‘Green Economy Strategy’ wing of CRGE.

In both the CRGE Vision and CRS-AF texts, despite mentions of the effects of climate change on people’s livelihood and social well-being, the economic growth-oriented motive becomes predominant. As such, it addresses the question why to adapt. In particular, the CRS-AF document states that ‘building resilience to avoid damage to our economy depends on understanding the threats and the priority areas for focusing adaptation efforts (FDRE 2015, p 3). Moreover, the document highlights the economic motive (see ibid, p 25) under the title of ‘framing the impacts of climate from an economic perspective’. Such framing not only fails to provide an equivalent social or political framing but also it presents the national economy in a socially aggregated manner using the term ‘our economy’. However, as feminist scholars point out, economic aggregates without reference to gender tend to cover the differentiated effects of policies on men and women (cf. Elson and Cagaty 2000), failing to acknowledge that they have a propensity to benefit men because of the gender division of labour that favours men and that fact that men have better access to productive resources (ibid).

Programmatic ideas: who or what adapts?

Given its origin, CRS-AF focuses on the agriculture sector and uses livelihood systems such as pastoral, agro-pastoral and cropping systems to define their exposure and vulnerability. As such, it pertains to the question who and what adapts. Adopting the definition set forth by the Intergovernmental Panel on Climate Change (IPCC), CRS-AF defines vulnerability as ‘the propensity or predisposition to be adversely affected’, where vulnerability is understood as a function of exposure, sensitivity and adaptive capacity (FDRE 2015, p 9). Biophysical exposure to climate change is used to analyse vulnerability, and it relies on quantitative and technical measurements to distinguish 18 climate stresses that affect different livelihoods and areas.

Except pointing out ‘women’, ‘disabled’ people and ‘children’ as being the ‘most vulnerable’ (ibid p, 28), the document fails to provide a qualitative analysis and socioeconomic assessments to reveal differences among and within communities. This is a typical example of how a dominant framing of women’s vulnerability leads to the uncritical inclusion of women as mere recipients of policy benefits (Bee et al. 2013; Skinner 2011). No explicit gender-based analysis is provided in an attempt to lift women’s vulnerability, instead. As a result, the underlying structural and institutional factors that put women in a vulnerable position remain undisputed. Gender relations and gender norms play pivotal roles in the lived experiences of male and female smallholder farmers in Ethiopia specifically with regard to unequal access to productive assets, extension services, agriculture inputs and overall masculine farming norms (cf. Aguilar et al. 2014; Buchy and Basaznew 2005; Gella and Tadele 2014; Mogues et al. 2009). This results in significant gender-based differences in farm productivity. For instance, Aguilar et al. (2014) find that male-managed plots produce on average 23.4% more per hectare than female-managed plots. Given these circumstances, adopting a vulnerability assessment that is not gender sensitive does not only cover the differences among and within communities but also restrains the effectiveness and relevance of adaptation options.

Policy ideas: how adaptation occurs?

The overall adaptation approach in the documents is centred on ‘no and low regret’ actions that intend to reduce vulnerability and enhance adaptive capacity (FDRE 2015). To identify 41 out of 350 suggested adaptation options, CRS-AF uses four criteria including relevance and feasibility; contribution to economic growth; equity and distributional issue; and balancing the present and future effects (ibid, p. 7 and 45). The 41 adaptation options are categorised under nine themes including capacity building and institutional coordination, information and awareness, crop and water management on-farm, livestock, value chain and market development, sustainable agriculture and land management, natural resources conservation and management, disaster risk reduction and social protection.

Gender- and poverty-related issues are dealt with by the equity and distributional criteria in the document (ibid). However, by equating gender with women, the document fails to consider gender as a structural and relational factor. As a result, women are singled out as a ‘high priority group’ under the theme of ‘social protection’ while the remaining eight themes and adaptation options under them are presented without any form of social disaggregation and in a non-gendered way.

From the document analysis, it becomes apparent that the framing of climate discourses and policy ideas at philosophical, programme and policies levels show congruence with global narratives and guidelines. As mentioned by several feminist scholars, the global discourses and narrations dominantly demonstrate a gender void (cf. MacGregor 2010).

It also tends to create a boundary between climate and non-climate triggers. The former was highlighted while the latter was systematically underplayed. As a result, the documents fail to explicitly address social dynamics and the differentiated impacts of both problems and solutions following gender lines, among other factors.

The next question is how and why such asocial and non-gender framing of climate policy discourses gained prominence, which will be discussed next.

Coordinative discourses: actors and interactions

Gendered actors: who is involved?

Recently, the climate change agenda has attracted different government actors, which, in some cases, has resulted in a conflict of interest. Especially since the 2009 United Nations Framework Convention on Climate Change (UNFCCC) Conference of the Parties (COP 15), significant changes have been made regarding national government actors’ mandates and the organisational structure. From the late 1990s to 2009, the Ethiopian Meteorological Authority (EMA), which was also a key player in the preparation of the National Adaptation Programme of Action (NAPA), was the focal point for the UNFCCC and the IPCC. However, after COP 15, the Ethiopian Environmental Protection Authority (EPA) took over this place and assumed a central position, both nationally and internationally.

The EPA was upgraded in 2013 to the Ministry of Environment and Forests, and then in October 2015 to the Ministry of Environment, Forest and Climate Change (MoEFCC). However, the organisational landscape and actors related to CRGE go beyond the MoEFCC and involve more national influential actors such as the Prime Minister’s Office and identified sector ministries that are represented in the Ministerial Steering Committee, which leads the CRGE (see Fig. 2). Also, the Ministry of Women and Children Affairs (MoWCA) is a part of this steering committee. In the government hierarchy of Ethiopia, the MoWCA and the Women’s Affairs Directorate (WAD) are the mandated actors to raise ‘gender’ or ‘women’s’ issues in policies and their implementation (Biseswar 2008; Østebø 2015).Footnote 11 The absence of a vibrant local civic society and feminist or gender advocates imparts additional responsibility to these women’s affair offices, as they are virtually the only actors operating in this field.

Fig. 2
figure 2

Structural arrangement and linkage of CRS-AF, DFID—UK’s Department for International Development, GGGI—Global Green Growth Institute, GCAP—Global Climate Adaptation Partnership, GeoSAS—Geospatial Analytical Services (local consultant), MoWCA—Ministry of Women and Children Affairs, MoEFCC—Ministry of Environment, Forest and Climate Change. Source: adopted and modified from FDRE (2015)

Full size image

The preparation of CRS-AF also followed the configuration of the CRGE hierarchy, but the MoANR played a key role. From that ministry, key actors were the departments of Agricultural Development, Natural Resources, Disaster Risk Management and Food Security, Women’s Affairs and the Agriculture Research Institute. Hence, much of the coordinative discourses were carried out among these departments, as well as the consulting teams (Global Climate Adaptation Partnership and a local consultant, Geospatial Analytical Services). In CRS-AF’s preparation, one expert represented MoANR/WAD. It was in the last few sessions that one additional expert from MoEFCC/WAD joined the discussion, as reported by interviewees.

Gendered interactions: how, when and why?

This part deals with the questions of how, when and why gender-related information is conveyed in the process. Since MoANR/WAD has the mandate to participate in the preparation of any policy, strategy or programme in the sector, it was invited to send one expert for the preparation of CRS-AF. However, the representation of WAD seems to have been merely a formality without being thoughtful of the issues that might arise from the department. One of the interviewees remarked that a gender-void viewpoint had emerged in the discussion during early meeting sessions.

When the staff of the global consultant team [GCAP] offered training, the presentation focused on the science of climate change, such as the concentration of greenhouse gases and how it affects energy transfer. Nothing was mentioned about gender. I asked the presenter why gender related discussions were not included, and he did not have any response. He simply said ‘We did not thoroughly think about it’.

In a similar account, another interviewee pointed out that the gender dimension was missing from the start because the Terms of Reference (TOR) themselves did not encompass any gender-related expectations. The TOR were prepared by the Global Green Growth Institute (GGGI) in collaboration with representatives of the MoANR (all represented by men), and neither MoANR/WAD nor MoEFCC/WAD were engaged. The absence of women’s voices or interests could have been one of the reasons for the observed lack in recognising that gender relations and women’s experiences are important. One of the verbatim records from the interview expounds as follows:

Regarding gender related questions, the consultants often appeared unprepared. In fact, holding the consultants accountable for something they had not agreed upon may not be fair.

Observing the repeated side-stepping of gender dimensions in the discussions and presentations, the WAD expert, with the help of others, prepared a text that elaborated the gender dimension of climate change in the agriculture sector and the rationale for its consideration.

After the short note was distributed, there seemed to emerge an acknowledgement on the need for considering gender. But then the dialogue shifted to considering gender at the project implementation levels instead of at the policy and strategy design levels.

This quotation clearly shows why the CRS-AF document fails to address gender explicitly, even after the concern was raised. Suggesting gender as a matter of implementation is a way of putting aside the agenda and systematically silencing the voices expressing gender concerns in the policy dialogue. This relegation of the gender dimension in policy making as a matter of implementation has been corroborated by Kronsell (2006) as well. This silence on gender implies that no discussion on the masculine construction of policy problems, methods of analysis and the proposed responses is effectuated (see also MacGregor 2010, 2014), resulting in policies in which the underlying gender inequalities and power imbalances continue to perpetuate.

It is also important to mention that WADs representatives who engage in the discussion are responsible to raise the overlooked gender dimension. As a consequence of Proclamation 601/2001, the engagement of Charities and Civic Society Organisations in policy advocacy is restricted in Ethiopia. Therefore, advocating gender issue falls entirely on the shoulders of WADs (see also Østebø 2015). However, our interviewees report that besides their limited number, the WAD representatives who are granted a seat at the table also have to deal with a limitation in terms of expertise in climate change-related discourses (see also Biseswar 2008)—an impression that is exemplified by the following quote from an interviewee:

There is no climate change expert in our department [Women Affairs Directorate]. I only took a one-day training on the issue. Thus, the knowledge gap also restrains us from having a meaningful engagement.

When the very actors who are supposed to carry on and push for the gender agenda encounter capacity limitations, it is easy to see why policy interaction would fail to go beyond a surface-level discussion in order to address structural and systematic problems.

Gendered institutions: formal and informal

WAD has the mandate to express and suggest gender considerations in organisational policies, programmes and practices. This formal institutional structure obligates ministries and departments to engage WAD experts in processes that lead to the framing of problems and the design of (policy) solutions. However, this requirement to invite WAD to the table when policies are discussed is of course not a guarantee for gender responsive outcomes. On the contrary, interviewees indicated that the engagement of Women’s Affairs officials in policy making or in daily organisational practices is highly influenced by gendered informal institutions such as social norms and informal networks despite the granted statutory equality.

Instead of discussing gender in detail, like why and how to consider it, discussions are often limited to where to put something about women. This is because gender is considered as ‘less scientific’ than other issues.

This quote elucidates the problematic conception of ‘gender’ as equivalent to ‘women’; hence, the mentioning of something about ‘women’ in the document is assumed as addressing gender. On a similar note, Biseswar (2008) argues that the core problem in advancing gender issues in Ethiopia is this narrow and arguably erroneous construction of gender and gender equality. Subsuming ‘gender’ under ‘women’ causes a detachment of women’s marginalisation from the root of the problem, and it needs to be seen as a function of power structures and institutional practices that accept and privilege masculine values (MacGregor 2010).

Drawing a further implication from the results of the interviews, it can be argued that the informal norm that trumps ‘hard’ climate science over ‘soft’ social science (and subsequently over gender) is another area of concern that hinders the emergence of gender-responsive policies and strategies.

As experts with a natural science background tend to dominate, discussions are often geared towards scientific facts and measurements. Although social dimensions are raised, they do not get equal weight and attention.

Another important issue that emerged from the interviews regards the problematic position of Women’s Affairs offices in the organisational hierarchy and politics. The respective positions they assumed undoubtedly create a constraint for their engagement in the policy dialogue as indicated in the following narrations.

After a long discussion, gender may end up being sidelined and we cannot do anything about it. Our responsibility is merely pointing out the issue. When the resistance persists, we just retreat to avoid confrontation.

Actors with a gender agenda [Women Affair Directorate representatives] should carefully choose their words and be systematic when they make comments because people become defensive and feel threatened easily, as they often overlook the gender component. And then this may negatively affect the working relationship.

These verbatim records highlight the challenges of negotiating a gender agenda for Women’s Affairs offices. As indicated by respondents, the policy making process, including how agendas are set, which issues gain more attention and the ‘scientific’ framing, carries a lot of gender baggage. By and large, the organisational power structure and hierarchy and masculine norms create an uneasy atmosphere for Women’s Affairs experts to extend their conversation. The last argument is in consonant with Acker (1992), who notes the influence of embedded androcentric values on organisational practices. Thus, despite the formal institutional setting in Ethiopia, which appears to be progressive and enabling, gendered informal institutions and organisational politics constrain policy processes and outcomes from becoming gender-responsive.

Discussion and conclusion

In spite of the growing recognition given to gendered vulnerability and adaptation, discussion on climate policy, discourse and institutions remains sporadic. This article aimed to explore the extent to which gender is represented in adaptation policy and how coordinative discourse (the interaction among policy actors) influences gender responsiveness in the crafting of adaptation options. Adopting a discursive institutionalist framework, the study analysed the case of Ethiopia’s Climate Resilient Strategy for Agriculture and Forest (CRS-AF) focusing on three dimensions, namely ideas (substantive content of the discourse), coordinative discourse and institutions.

The policy document analysis revealed an overall failure of CRS-AF to explicitly address gender dimensions. The sheer mention of ‘gender’ and ‘women’ in some instances does not equal true gender responsiveness. Our analysis of the data on adaptation to what and why (philosophical ideas) shows that the document mainly deals with the issue of climate change from a natural science perspective—specifically, it contains details on temperature and rainfall change, climate variability and extreme events such as drought and flooding. At this level, the text avoids mentioning the gender dimension and barely refers to social aspects related to climate change. Emphasis is given to the perception of climate change as an exogenous phenomenon by minimising the role of non-climate drivers. Our analysis also indicates that CRS-AF tends to be highly influenced by an economic motive that is lacking in gender (and social) equity issues.

As regards the question of who and what adapts (programmatic ideas), the usage of biophysical vulnerability assessment that only pays attention to physical damage obfuscates the different levels and experiences of vulnerability across groups in the society. The document cites women, people with disabilities and children as the most vulnerable, but with no equivalent analysis of how and why these different groups were included in this category. This grouping not only reflects an essentialist assumption about women but is also paternalistic, as the root causes of women’s vulnerability remain largely concealed.

Regarding the question of how adaptation occurs (policy idea), as the vulnerability assessment did not consider gendered power relations, institutions and other socio-economic drivers, gender-responsive adaptation options are hardly considered at this level. As the agriculture sector in Ethiopia is highly influenced by gendered social norms, and most policies and services are criticised as being gender-biased (Buchy and Basaznew 2005; Mogues et al. 2009), adopting ongoing practices as adaptation options without a substantial and critical analysis of the gender dimension can exacerbate the existing gender inequalities (Mersha and Van Laerhoven 2018).

In general, we found that gender-neutral wording dominates the CRS-AF documents. According to gender and feminist scholars, such construction of policies and strategies often reinforces and reproduces masculine values (Bee et al. 2013; Arora-Jonsson 2011; Pearse 2016). By doing so, this not only preserves the status quo but also naturalises gender inequality (Kronsell 2006). The subsequent question is thus why new policy and strategy development remains silent about the gender dimension of climate change impact and responses under a cover of gender neutral language even in an era that witnesses a growing debate around the recognition of the gender dimension of climate change. To begin answering this question, our analysis of the coordinative discourse allowed us to trace how specific policy ideas were constructed, justified and, finally, naturalised in the policy making process. Our analysis focused on who was involved, how ideas related to gender were raised and when and why they failed. In doing so, the following four issues emerged.

Numerical imbalances

The limited number of actors with an explicit gender agenda participating in the coordinative discourse made it difficult to create an alternative, gender-based discourse. One may argue that merely having more actors with a gender agenda would not necessarily guarantee bringing gender to the forefront of policy dialogues, as studies report the limited effect of women representatives in ensuring substantive contribution (cf. Holvoet and Inberg 2014; Magnusdottir and Kronsell 2015). Yet, on top of merely guaranteeing a seat to women representatives, what is equally important is ensuring that they have an explicit and well-crafted gender agenda, so they can make a substantive contribution. As alternative discourses need actors who can carry them (cf. Wodak and Meyer 2001), our suggestion is to include more actors with an explicit gender agenda.

Timing

In line with Skinner (2011), our analysis shows that gender ideas entered the coordinative discourse as an add-on concern at a relatively late stage. Lack of WAD and other gender experts’ participation in the development of the TORs demonstrates a lack of full comprehension of a more gendered approach. Consideration of the gender dimension has been sporadic, and the engagement of experts in the area was found to be a matter of formality. Consequently, when alternative ideas related to the gender dimension of climate change impacts and responses arose, they were easily sidelined or treated as a mere matter of implementation.

Natural science bias

Our document and interview analysis revealed a disproportional emphasis on the ‘scientific’ construction of climate change. According to MacGregor (2014, p 627), such scientific framing obscures power relations ‘that have caused, and are standing in the way of addressing, the current predicament’. Moreover, the resulting simplification of policy frames became a powerful force that pushed aside competing frames based on more complex ideas about climate change, including ideas that acknowledge gender dimensions.

Gendered norms and organisational practices

We found that gendered informal institutions serve as settings through which the issues mentioned above are enacted. It is also worthy to posit the role of gender prejudices at local level if they play such a role at national level. Thus, in a society where androcentric values remain an inherent part of formal and informal organisational and institutional structures, policy and strategy documents need to provide an explicit outlook on gender equality. Effecting a gender-sensitive framework can be a tool for regional and local Women’s Affairs offices to hold implementing bodies accountable. From our analysis, the main insights that can be drawn include:

  • The necessity of considering a gender agenda from the very initial stages such as the development of TORs and the importance of having a sufficient number of capable actors to represent the gender dimension in coordinative discourses.

  • A clear understanding of what gender responsive policies entail. It does not mean adding on the term ‘gender’ or the inclusion of women as ‘target groups’. Rather, it means recognising and addressing the underlying asymmetric gender relation manifested in different ways such as resource regimes, institutions (formal and informal) and other structural impediments.

  • To effectively address gender in the policymaking process, understanding and redressing the gendered informal institutional arrangement is at least equally important as merely guaranteeing a seat for actors with gender interest.

  • The explicit recognition of gender as a power relation in climate-problem construction, vulnerability analysis and development of adaptation options. The absence of a real commitment and the relegation of gender as a matter of implementation not only reinforce the socio-economic and power dominance of men but also, as MacGregor (2010) states, prevents adaptation options from being effective and sustainable.

  • Last but not least, we would like to highlight that the DI approach can be a powerful tool for the analysis of other social categories such as ethnicity, age, localities and livelihoods, as well as their interplay. These social dimensions of climate change policies could be the subject of future research.