How to diagnose institutional conditions conducive to inter-sectoral food security policies? The example of Burkina Faso

Highlights

•	The multi-dimensionality of food security requires inter-sectoral policies.
•	Food security policies are usually food production-oriented in African contexts.
•	Institutional diagnostic of food security policies is needed before prescription.
•	Visions, measurements and forums on food security are institutions to be diagnosed.

Abstract

The multidimensional nature of food security often leads experts to recommend mobilising all public intervention sectors to ensure that food security policies are inter-sectoral, and not the sole responsibility of a single sector. However, in African contexts such as in Burkina Faso, food security policies are in most cases far from being inter-sectoral. They are instead focused on agricultural production. It is therefore critical to understand why food security policies are what they are, to identify the underlying sectoral logics and to seek for signals of policy changes. This paper aims at contributing methodologically to the literature focusing on institutional diagnostic of food security policies. Drawing on a combination of new institutional approaches and cognitive public policy analysis we explain food security policies in Burkina Faso by three major factors. First, the persistence of agricultural production-oriented policies points to path dependency arising from the way food insecurity has historically been framed around cereal deficits. Second, the instruments used to measure and assess food security are not neutral: they directly shape both policy debates and decision-making. Third, the institutional configuration of the policy debate is characterised by a fragmentation that influences power games between actors supporting different visions of food security. Finally we argue that new concepts such as “nutrition-sensitive agriculture” combined with more open forums may have the potential to lead to more inter-sectoral food security policies.

Keywords:

• Food security
• Public policy
• Intersectorality
• Food production
• Institutional diagnostic
• Burkina Faso

1 Introduction

The multidimensional nature of food security is now firmly recognised¹ and at the same time poses serious challenges for public action. The idea that food security governance should mobilise all public interventions sectors (agriculture, employment, health, environment, trade, etc.) to address the complexity of the issue appropriately is often advocated by food security actors.² Many terms are used to express this idea – breaking silos, integrated, comprehensive, cross-sectoral, multi-sectoral or inter-sectoral approach, etc. – which we propose to capture under the term “inter-sectorality”. The advocacy of inter-sectorality often leads to the recommendation to institutional reforms to anchor food security under the Prime minister or the Presidency (IEH, 2012). Better inter-sectoral coordination is seen as a guarantee to avoid food security falling under the remit of a single sector and to ensure a well-balanced approach to its four major dimensions (access to food, food availability, utilisation and stability).

However, observations on the ground in African contexts show that food security policies are quite far from inter-sectorality. They tend to be primarily focused on agricultural production, obscuring other important dimensions such as economic access to food and nutrition.³ The Burkina Faso example is particularly illustrative of this bias. In this country, a third of the population (35.4%) is food insecure and this number keeps increasing. The dramatic increase in cereal production over the last four decades has not prevented food surplus areas – such as the Boucle du Mouhoun region which is the “grain basket” of the country – from being free of food insecurity and malnutrition. Paradoxically, food security policies hardly consider this reality and tend to focus on the single message of increasing food production. So, why are food security policies so production oriented in Burkina Faso despite the official rhetoric of the need for inter-sectorality?

This paper addresses specific institutional concerns of food security policies in Burkina Faso. Our core research question is how to diagnose institutional conditions conducive to inter-sectoral food security policies? It is indeed critically important to understand why food security policies are what they are in Burkina Faso and to identify elements that could be signals of possible policy changes before envisaging any institutional reforms. Analysing major development strategies of the last 50 years, Rodrik (2010) advocates for “diagnostics before prescription” to make sure these strategies are adapted to specific realities (i.e. locally suited and with remedies that vary over time as the context changes). Failures to achieve progress in development arise when models that are valid only in specific circumstances are transformed into universal remedies, while successes often result from pragmatic experiments (policy innovation and learning). In particular, institutions are key parameters to consider in any diagnostic, alongside others such as biophysical or economic conditions, before implementing scientific, technical or political solutions to address food insecurity in the African local context.

As noted by Stone (1992), institutions are defined very broadly by political scientists, mainly as long-term rules of behaviour, principles, norms and complex symbolic, discursive structures and rituals. These “structures” (institutions) can be material (e.g. a particular organisation) or non-material (e.g. a particular ideology), formal (e.g. a legal rule) or informal (with no coercive power). All these norms and principles can influence actors’ behaviours – serving as constraints to limit room for manoeuvre – but also to legitimise particular behaviours and de-legitimise others while providing opportunities for actors. Economists also underline that “institutions both constrain and enable behaviour” (Hodgson, 2006). Far from being only constraints for individuals, institutions are the outcomes of human interactions (North, 1991). This justifies paying attention to both agency and institutional structures.

A number of neo-institutionalist approaches in political science provide useful insights to analyse public policies, and especially to understand institutional lock-ins and processes of policy change. In particular, the historical and discursive branches of neo-institutionalism highlight the role of institutions to explain public policies (e.g. Pierson, 2000; Fouilleux, 2004; Schmidt, 2008). However, this stream of literature is mainly concerned with public policies in developed countries and rarely deals with developing country issues, such as food security in Africa. On the other hand, there is an increasing literature stream on food security and nutrition governance which especially addresses the issue of inter-sectorality (Reich and Balarajan, 2012; Garrett and Natalicchio, 2011; IEH, 2012). However, this literature is developed primarily with an operational objective, provides little theoretical ground and often lacks historical depth.

In this paper, we draw on new institutionalist approaches to analyse food security policies in Burkina Faso. We then propose a framework for the institutional diagnostic of food security policies in African contexts. We argue that food security policies in Burkina Faso can be explained through three factors which are key components of an institutional diagnostic: i) path dependency around the initial framing of food security and agricultural sectoral logics built over time; ii) the measurements, statistics and instruments of food security as micro-institutions that frame policy debate and decision-making; iii) the power games of actors supporting different visions of food security in the different forums where food security is discussed.

In Section 2, we present our theoretical framework. Data collection methods are in Section 3. Our results, which explain why food security policies in Burkina Faso are so production-focused, are presented in Section 4, and discussed in Section 5 in view of proposing a framework for an institutional diagnostic of food security policies. Section 6 draws some conclusions.

2 Theoretical framework: institutions in the analysis of public policies

Our theoretical framework to explore the production bias of food security policies is based on the combination of three perspectives on institutions (for an analysis at the global level, see Fouilleux et al., 2017). First, we draw on historical institutionalism, which provides useful insights to understand continuities in public policies and why they are “change resistant”. Second, we consider the stream in political science that focuses on “ideas”, i.e. non-material or cognitive structures, to explain policies. The so-called French school of cognitive analysis of public policy can be attached to the discursive branch of neo-institutionalism (Schmidt, 2008). Third, we use the literature considering measurements, statistics and instruments as another variable shaping policy debates and decisions.

2.1 Path dependency or the institutional lock-in of public policies

One major contribution of historical institutionalism has been to emphasize the weight of past policy choices on on-going policy definition processes. In such an approach, public policies tend to be reproduced independently of their relevance. Inspired by David (1985) and his work on technological lock-in, Pierson (1993, 1994) has taken up the concept of path dependence to analyse forms of institutional lock-in of social policies. According to this concept, once one type of public policy has been adopted, change becomes difficult because it is too costly. Pierson (2000) borrows the notion of increasing returns from the economists to not only explain that history matters but also that a particular path of action is increasingly difficult to exit over time. Policies are costly to create and often generate learning and coordination effects. The development of actors’ interests around the policy also makes the initial choice become irreversible and the policy become locked onto a given path. This view is summarised by Hall and Taylor (1996: 941) as “past lines of policy [will] condition subsequent policy especially by encouraging societal forces to organise along some lines rather than others (…) or to develop interests in policies that are costly to shift”.

However, the difficulty of policy change may depend on the degree of institutionalisation of public policies. Indeed, public policies are highly or weakly institutionalised depending on the sector. In the example of HIV policy, Lascoumes and Le Galès (2007) show that this emerging sector of public action is characterised by a weakly institutionalised policy. The representations of the public problem are not really stabilised, competing visions exist and there is no consensual definition of the problem because the definition process itself is an object of power struggle. On the contrary, other public policies are highly institutionalised as they have existed for a long time; actors involved in the sector are well organised within the administration and are well structured as interest groups with enough resources to defend their interests and impose their own problematisation of the sector. The ministries involved in such policies are often powerful, sometimes organised around large state bodies, relations between state and civil society are stabilised, the rules of the game are clear and stabilised, etc. Practices are often conducted as a matter of routine which are recognised as legitimate, and hence not questioned, as if they were set in stone. Corresponding policies are thus much more difficult to change than the former. Sociology and anthropology pay a particular attention to institutions such as formal or informal procedures, routines, symbols and cognitive scripts. For Olivier de Sardan (2013: 291), all informal norms visible in sets of routines, practices, habits, classifications or procedures are “practical modes of path dependency”. As developed by Nelson and Sampat (2001) to characterise social technologies, the ‘routine concept’ refers to “a collection of procedures which, taken together, result in a predictable and specifiable outcome” (p. 42). However, although routines are more or less automatic activities, the authors stress the fact that they do not eliminate choice and that there is room for variation within a routine.

While particularly enlightening for understanding why public policies persist over time, historical institutionalism gives less attention to the role of ideas and the interplay of actors in policy processes. This leads us to the cognitive analysis which provides a complementary view to capture informal institutions.

2.2 The cognitive analysis of public policy or “ideas in action”

A number of political scientists have put forward “ideas” to explain public policy processes. Schmidt (2008) characterises this “turn to ideas and discourse” as a fourth new institutionalism which she calls “discursive institutionalism” and positions after the three new institutionalisms distinguished by Hall and Taylor (1996), namely, rational choice, historical and sociological institutionalism. By taking ideas and discourse seriously, this discursive institutionalism offers, according to the author, a more dynamic view of policy change than the first three forms of institutionalism. These first three forms tend to leave aside agency and to see institutional rules only as constraints because institutions are treated as a given and external to actors: “subordination of agency (action) to structure (rules)”. The singularity of discursive institutionalism is to treat institutions both as a given (structures) and as contingent or constructs created and changed by actors. This fourth approach claims therefore to be an “agent-centred approach” and to put “agency” at the heart of the explanation of institutional change (or continuity) (Schmidt, 2008).

Schmidt distinguishes between two types of ideas: “cognitive ideas” serve to justify policies and refer to interests, while “normative ideas” attach values to political action and serve to legitimise policies. Discourse is “not only the substantive content of ideas or text (what is said) but also the interactive processes by which ideas are conveyed (where, when, how and why, who said what to whom)” (p. 305). Policy actors – i.e. individuals and groups who are at the centre of policy construction – seek to coordinate themselves on policy ideas on the basis of shared cognitive and normative ideas, thus developing a “coordinative discourse”. In addition, they try to convince and gather other actors around their position through a “communicative discourse”. In this paper, we focus mainly on coordinative discourses.

Older than the discursive institutionalism, the cognitive analysis of public policy (or so-called French school of public policy analysis) also highlights the role of values or symbolic constructions that influence actors’ representations and behaviours in policymaking. It describes sectoral public policies through the concept of “referential” (Jobert and Muller, 1987; Jobert, 1994, 2004). As a system of representations of the world and of public problems, the referential embraces both a cognitive dimension (how a problem is understood and interpreted) and a normative dimension (what solutions should be implemented). These two dimensions correspond to two different tasks in policy design, the former being a decoding of reality and the latter a recoding of reality. Changes in sectoral policies are then explained by adjustments between sectoral and global referentials. The global referential describes the representation of the world associated to a set of core values and beliefs in a given society (a nation), while the sectoral referential refers to the dominant representation of the problem in a given sector (Muller, 1990). When global visions change, sectoral policies are “constrained” to change. The task to make the sectoral referential match the new global referential is operated by specific actors – so-called “mediators”.

In this analytic tradition, the sector is not limited to an institutional setting such as a ministry or a department within an organisation. Instead, it refers to a way of organising social actors, forms of mediation between these actors and to public policies relying on a specific vision of the public problem and on specific instruments for implementation (Muller, 1990). By emphasizing that sectoral referentials are social constructs resulting from the action of actors sharing the same visions of the world (and of the sector), cognitive analysis of public policy also puts agency at the heart of policy change. Sectoral referentials do not impose themselves as legitimate if certain groups of actors do not struggle for a dominant position within strategic power spaces, the “mediators” being particularly active by using specific technical, political and discursive resources (“ideas in action”).⁴ In other terms, the mediators act as institutional entrepreneurs.

Beyond the global-sectoral adjustment process, other authors in the same school of thought have developed analytical tools to better understand the complexity of policy change. They insist on the importance of the institutional configuration of spaces for policy debate and distinguished between forums and arenas (Jobert, 1994; Fouilleux, 2000, 2004; Fouilleux and Jobert, 2017). Policy forums are spaces where competing policy ideas are confronted and selected. Arenas are spaces where compromises are negotiated and policy instruments are decided. Therefore, these two kinds of institutions do not have the same influence and ability to shape the debate. In the words of Stone (1992), this approach analyses the interactions between material (institutional configuration of spaces for policy debate) and non-material or cognitive structures (referential), the former providing means for expressing the latter. Authors also argue that the situation of a dominant referential does not prevent controversies; any referential is constantly questioned and destabilised and when being relayed on different forums and fuelled by large social mobilisation, those controversies may lead to radical policy changes (Fouilleux and Jobert, 2017).

The trajectory of any particular idea (emergence, imposition, persistence) is strongly contingent on the institutional configuration of spaces for policy debate, which leads us to also consider how ideas can circulate across different forums (professional, scientific, national, global, etc.). In the African context, some political scientists have studied policy transfers through the lens of ideas and discourses. They show that “mimetism” of external policy models are nevertheless always adapted, re-appropriated and reinterpreted: see for example the work of Darbon (2004) on African public policies embedded in complex dependency relations with Northern countries, Balié and Fouilleux (2008) on West African common agricultural policies inspired by the European one or Bergamashi (2001) who studied the trajectories of words and instruments related to the post-Washington consensus in Mali.

2.3 Measurement instruments as micro-institutions

Public policies do not only reflect broad objectives and political choices but they also translate into concrete instruments. Beyond their technicity, policy instruments are far from neutral since they convey specific values, political choices and the problematisation of the object of public action (Lascoumes and Le Galès, 2007). Instruments are therefore institutions that strongly influence the framing of policy debates and decision-making; they are “micro-institutions” compared to public or macro institutions (Fouilleux, 2000; Fouilleux et al., 2017). Statistics are specific micro-institutions that reflect representations of public problems and are used by actors to impose their own representations (Desrosières, 1993/2000Desrosières, 1993/2000). While statistics are necessary as a language to facilitate contradictory debate on policy options, Desrosières states that this language is the result of a complex history of practices which are negotiated, stabilised, then distorted and questioned. Ilcan and Phillips (2003) also show how statistical practices and census knowledge promoted and used by the FAO aim to “not only render social domains calculable” but to “feed global policies, such as those of multilateral trade” (p. 456). The choice of data, indicators, statistical methods, etc. that are used to characterise and measure public problems therefore constitute institutions that are important to study in the analysis of public policy.

In short, the theoretical framework adopted in this study focuses on three main parameters through which an institutional lens can be operationalised to understand food security policies: (i) path dependency in actors’ representations of the public problem and of its solutions, (ii) the instruments used to characterise and measure the problem and (iii) the power games between actors supporting different views of the problem and of its solutions. Under this perspective, our research questions are: how does the specific representation of food security through the main lens of food production emerge? To what extent do data and statistics on food security contribute to such representation? To what extent does the institutional configuration of debates explain the state of play of the bargaining power of different actors?

3 Methods

This paper is mainly based on a CIRAD political science research programme dedicated to the analysis of food security policies in West Africa, for which the first author has been based in Burkina Faso for three years. This long term presence provides an exceptional opportunity to build trust relations with some actors involved in policy processes, to participate in development projects and provide expertise on a range of food security issues (resilience, agriculture and nutrition, market access, etc.). More specifically, various methods have been implemented to gather the empirical evidence used in this paper.

Firstly, the key actors involved in the definition and implementation of the national food and nutrition security policy were identified, in the administration, civil society and aid sector (donors, international organisations and NGOs). More than 50 semi-structured interviews were conducted with the different categories of stakeholders in the administration including decentralised administrations and research institutions (n = 30,), in civil society organisations including producers organisations (n = 6) and in the aid sector (n = 20). The purpose of the interviews was to collect information on their visions, interests and arguments on the food security problem and its solutions by asking questions such as: what does food security mean, how can we characterise and measure it, what are the issues at stake, what solutions should be promoted, what interventions are implemented, what is the role in the policy-making process, the view on the institutional configuration of debates, etc.

Secondly, some interviews with administrative officers from statistic services and early warning system involved in the food security information systems (n = 4) were instrumental to collect information regarding the different measurement instruments of food security.

Thirdly, the key forums where food security issues are discussed were identified. Attending meetings in those forums allowed us to observe the participants, their relationship, the agenda, discussions, controversies, negotiation of compromises, etc. Regular participatory observation has been possible in the National Council on Food Security (Conseil National de Sécurité Alimentaire, CNSA), which is the reference body in charge of food security in Burkina Faso, and in the aid actors’ group on food security and nutrition.

Finally, discourses were collected through an extensive review of documentation, both from policy documents and key projects from government and/or from donors and NGOs. A reading grid was constructed in order to identify the vision of food security that emerges from the documents, how it is characterised, assessed and what policy options are explicitly or implicitly chosen. Documents were then qualitatively analysed through this reading grid.

4 Why is it so hard to develop inter-sectoral food security policies in Burkina Faso?

The analysis of all materials collected on food security policies in Burkina Faso allows us to show a predominant bias towards food-production and to describe three main underlying factors: i) the weight of history in the way food insecurity and its solutions have been conceived (path dependency in framings); ii) the administrative routines created by this historical weight in the measurement of food security (instruments); and iii) the interplay between historical agricultural actors and other actors supporting alternative visions to production (power games in forums). This institutional lens composed of three parameters – ideas, instruments, forums – also allows us to identify changes that could constitute some fundamental premise of policy reforms.

4.1 Facts about food insecurity and agricultural production-oriented food security policies

Despite sustained economic growth and dynamic cereal production over the last decade (respectively 6% on average per year between 2000 and 2012 and 4.4% on average per year from 2004 to 2014), food insecurity continues to grow in Burkina Faso (Burkina Faso, 2014a; FAOSTAT). According to the last national survey on food security (MAHRH, 2009), one household out of three (35.4%) is food insecure, especially in rural areas where the proportion is higher (37%) than in urban areas (31%). While this situation may sometimes be due to insufficient cereal production, food insecure people are generally poor producers who sell their cereals to get cash and then buy cereals in the lean season when granaries are empty and prices are higher. Therefore, food insecurity exists even during good cropping seasons because of credit problems and insufficient incomes generated by cash crops.

Significant progress has been achieved in reducing undernutrition and malnutrition over the period 2003–2012 (due to nutrition-specific interventions such as food fortification) but, prevalence rates remain high: stunting prevalence for children under 5 decreased from 38.7% to 32.9%, wasting from 19% to 10.9% and underweight from 38% to 24.4% (Burkina Faso, 2014a). Micronutrient deficiencies (hidden malnutrition) also represent a major problem: 88% of children under 5 suffered from anaemia (iron deficiency) in 2010 (ECOWAS-CAADP, 2011). A number of issues are at stake: consumption practices, education, hygiene, women’s workload, childcare, etc.

The “paradox” is that food surplus areas are far from being exempt from food and nutrition insecurity. The Boucle du Mouhoun region is emblematic of this paradox: located in the western part of the country and considered to be its grain basket, the region has levels of poverty, food insecurity and infant malnutrition above national averages (INSD, 2015).⁵ Different factors combine to account for this paradox: low yields for some poor producers but also factors such as marketing, prices, incomes or poor food diets. In short, such a “paradox” simply illustrates that food production is not enough to ensure food security and that one should pay attention to many other factors for a good understanding of the situation.

The Burkina Faso National Policy on Food and Nutrition Security (Politique Nationale de Sécurité Alimentaire et Nutritionnelle, PNSAN) published in October 2014⁶ underlines the cross-cutting nature of food insecurity and the need for better inter-sectoral collaboration.⁷ The document presents five strategic areas through which to achieve sustainable food and nutrition security by 2025 and which cover the four pillars of food security.⁸ However, the analysis of their content shows that much higher importance is given to issues of agricultural production (including animal resources, fisheries, forestry, etc.) compared to access to food and nutrition. The three-year rolling 2014–2016 plan of action finalised in October 2014 is also illustrative, with half of the budget allocated to boosting agricultural productivity (Burkina Faso, 2014b).

Flagship governmental projects, programmes and instruments dedicated to food security are mainly in the hands of the Ministry of Agriculture (MoA).⁹ Food security and sovereignty was the first component (out of five) of the 2011–2015 National Programme for the Rural Sector (Programme National du Secteur Rural, PNSR), which operationalised the previous food security policy document, the 2003 National Food Security Strategy (Stratégie Nationale de Sécurité Alimentaire, SNSA).¹⁰ It focuses mainly on increasing agricultural productivity and competitiveness, only assigning a sub-component to the prevention and management of food crises. In line with the PNSR, the Programme for the Improvement of Food Security (Programme d’Amélioration de la Sécurité Alimentaire, PAPSA) corresponds to agricultural interventions such as infrastructure in lowland areas to fight against erosion and manure pits.

The same observations that food security interventions usually equate with agricultural ones and hardly integrate nutrition and social interventions can be made regarding the practices of donors and NGOs. One illustration is the formulation of the first sectoral budgetary support from the European Union in the field of food and nutrition security, sustainable agriculture and resilience (Programme d’Appui à la Sécurité Alimentaire et Nutritionnelle, à l’Agriculture Durable et à la Résilience, PA-SANAD). Three priority sub-sectors were identified – small irrigation, small livestock and nutrition – the first two are considered as the “food security sub-sector” and benefit from agricultural budget support, whereas nutrition is considered separately with a different funding modality, which is through NGOs.

This situation of agricultural production-oriented food security policies creates problems. Other dimensions, such as those related to nutrition and poverty lack sufficient consideration. The diversity of food security situations also receives inadequate attention: production interventions tend to be the unique option envisaged whatever the characteristics of the areas. Such limitations do not allow complex situations to be properly addressed, such as in the Boucle du Mouhoun region, where food insecurity and malnutrition coexist with food surpluses. So how can we explain such a situation that is disconnected with the discourse on inter-sectorality and with the reality of food insecurity?

4.2 Historical framing of food security: the tragic food crises of the 1970s

The fight against hunger has always been a concern for village communities who have organised their own food supply through very dense regional trade networks and collective granaries, but it became a “public problem” when the colonial state developed. In the 1930s, regional trade was profoundly transformed so that trade was mostly conducted with the colonial state, disrupting the supply through traditional commercial routes to food-deficit areas, creating local food shortages. The colonial administration then established famine granaries that could only be opened by order of the governor in case of major food crises (Poussart-Vannier, 2006).

Food crises linked to severe droughts in 1973/74 marked a turning point in the public management of food security in Burkina Faso.¹¹ The public agency that was created in 1971 for the regulation of the cereal market, the National Office of Cereals (Office national des céréales, OFNACER) shifted its focus to the management of food aid. In a context of structural deficits in cereal production, an increase in food production was necessary to avoid other major famines. Cereal banks were developed throughout the country to manage cereal surpluses in case of food shortages. In the mid-1980s, under the revolutionary presidency of Sankara, a strong agricultural policy was implemented to achieve food self-sufficiency through increasing local agricultural production. Over the whole period, food crises were seen as a result of insufficient cereal production and the solution was therefore to support agricultural producers.

The liberal orientation in the 1990s and the structural adjustment reforms implemented in the country led to significant change in the representation of food security. A new policy discourse from the international financial institutions (the Washington consensus) was imposed on food security. Access to food, be it locally produced or imported, was emphasized and economic support to the agricultural sector declined. The institutional framework for the prevention and management of food crises, still in place today, was built at this time with the creation of the National Company for Food Security Stocks Management (Société Nationale de Gestion du Stock de Sécurité alimentaire, SONAGESS) to replace OFNACER, the Early Warning System (EWS) and the National Committee for Emergency Relief and Rehabilitation (Comité National de Secours d’Urgence et de Réhabilitation, CONASUR). Under this framework, food deficits are predicte and cereals are transferred from surplus to deficit areas and distributed to targeted population.

Although food insecurity was strongly associated with food vulnerability and food assistance, the issue continued to be viewed through the lens of agriculture. Food security was part of the Agricultural Sector Adjustment Programme (Programme d’Ajustement Sectoriel Agricole, PASA) and the food security institutional framework was named the Committee of Reflection and Monitoring of Cereal Policy (Comité de Réflexion et de Suivi de la Politique Céréalière, CRSPC). Beyond changes in the policy discourse, institutional practices and related instruments thereby became embedded in the agricultural sector.

The first reference policy document on food security, the 2003 SNSA, aimed to achieve its global objective of reducing by half the number of food insecure people by 2010, primarily through the specific objective of increasing food production. No significant change in the dominant representation of the problem of food security and of its solutions was brought to the 2014 PNSAN, even though the problem of food insecurity is much more complex now than in the 1970s. Indeed, the tragic events of the 1970s made a profound mark on the collective memory, which today continues to frame food security representations and to shape policies.

4.3 The food security information system to prevent and manage food deficits

Contrastingly, the food security information system (Système d’Information sur la Sécurité Alimentaire, SISA) has significantly evolved during the last decade to embrace different dimensions of food security. However, the core information used for decision-making remains consistent with the initial representation of food security.

The first information systems on food security were built in the 1980s in a context of food deficits. Their objective was to monitor cereal production so that the alarm could be raise in the event of any major food shortage and emergency responses could then be prepared in the framework of the CRSPC. Today, the SISA coordinates different sectoral information systems produced by several ministries. It is the result of successive expansions to cover the main aspects of food security. Indeed, the Plan of Action of SISA (PA-SISA) adopted in 2004 was very critical of the system’s focus on food production and its inability to be in line with new knowledge on food security.¹² Hence, the SISA has integrated more and more components: food balance sheets instead of cereal ones, the Household Economic Analysis (HEA), the poverty profile produced by the National Statistical Office through extensive surveys on household living conditions, nutrition and food vulnerability in urban areas. One interviewee from the MoA explains: “gaps are progressively filled. Food security assessment was limited for a long time to food availability. When we were talking about access we were only doing a brief price analysis”.

However, the progressive expansion of the SISA also rendered its use more sophisticated and complex. None of the sectoral information systems produces information entirely dedicated to food security. It is the cross-analysis of the different information systems which is supposed to provide the food security situation. The multidisciplinary working group in charge of the coordination of the sectoral information systems is confronted with the technical difficulty of doing such analyses, not to mention all problems of staff and capacity that prevent the group from optimally functioning.

In the daily functioning of the CNSA where all actors involved in food security are supposed to debate and decide on food security interventions, the discussions are in fact predominantly based on farming season information, which is mainly provided by the agricultural statistics services of the MoA. Data on food security therefore essentially consists of the EWS, which tracks changes in cultivated land, stocks, migration and prices. Nutritional data are also considered but are generally juxtaposed with production data, thereby not allowing for deeper analysis.¹³

The long experience and expertise built up over decades by the administration in monitoring farming seasons thus created strong routines in administrative practices in the way information is produced and policy debate on food security is organised. This leads to some confusion between the monitoring of food security and that of farming seasons and therefore food security policy tends to be a “seasonal policy” focused on the lean season.¹⁴

4.4 Power games in forums: the legitimacy of agricultural actors as regards food deficits

The long-standing idea of achieving food security mainly through agricultural production does not mean there have been no alternative visions contesting this idea over the years, but rather that it results from the action of specific actors in forums where food security policy debate takes place. The actors historically at the forefront of bringing solutions to food insecurity, which was diagnosed in the 1970s as a problem of cereal deficits, were officials from the agricultural administration, producers and agricultural experts. Today, the profile of actors involved in the food security field remains predominantly attached to the agricultural sector. Most actors are agronomists, livestock engineers or agricultural economists; relief actors seem to have a more heterogeneous profile, in geography or law for instance. This type of academic background and skills for food security actors is decisive in framing their vision of food insecurity around agricultural issues and the policy options considered.

Therefore, the strength of the idea first relies on the predominance of agricultural actors in the food security field who benefit from an historical legitimacy compared to non-agricultural actors and from a more powerful ministry with a larger budget. The legitimacy of agricultural actors is also built on the fact that people suffering the most from food insecurity live in rural areas and work in agriculture. In addition, agricultural producers have been organised for longer than consumers for instance, and they therefore represent a social force with a particularly strong legitimacy with respect to food security.

Secondly, agricultural actors – like all other actors involved – fight to impose a vision of food security that supports their own interests and maintains their leadership in food security debates. The dominant vision of food security around agricultural production is not “a given” but has resulted from permanent power struggles. For example, following the move towards neo-liberalism that led to the emergence of a vision of food security that was more reliant on international markets, competing visions such as food sovereignty appeared in global forums and in Burkina Faso. Producer organisations – members of the Via Campesina international movement through the Network of Farmers’ and Agricultural Producers’ Organisations of West Africa (ROPPA) – support food sovereignty as a political approach to defend local family farming opposed to the “too technical” food security approach. After the 2008 food crisis, food sovereignty became an explicit objective of the government – no more an indicative term from civil society – to justify the priority given to local production to decrease the country’s exposure to international price volatility and to reduce the food import bill. Different actors actively pursuing their own interests and using different arguments thus converge to promote the same idea of increasing agricultural production.¹⁵

The drafting process of the PNSAN is another illustration of agricultural actors’ action in shaping the content of food security policy. A multisectoral working group of around 15 technical executives from agriculture, animal resources, environment, trade, finances, transport, etc. was created but it was spearheaded by a previous Minister of Agriculture and the whole process was led by the Executive Secretary of the CNSA under the MoA. From the civil society side, the National Federation of Producer Organisations was particularly active in promoting production issues. The drafting process remained therefore agriculture-driven and the PNSAN – though globally considered satisfactory by most actors – reflects the vision of agricultural actors. This is well summarised by one interviewee from the aid sector stating that: “The document is led by agriculture, which is not surprising to see as it’s an agricultural document. The MoA tends to only describe production or even exportation results, but not food insecurity. Food insecurity is not part of its concern! We didn’t succeed in including a single chapter on food security”.

Thirdly, the CNSA, which was created in 2003 to steer food security policies, presents itself as an inter-sectoral body with the mandate to convoke all relevant ministries, but it is largely controlled by agricultural actors and is institutionally anchored under the MoA. Some interviewees stated that the MoA has no difficulty calling on representatives of sectoral ministries and is widely open to all stakeholders. Other interviewees underlined that the MoA is much less familiar with the ministries of social affairs or health than with the ministry of animal resources for example. Indeed, social affairs and health representatives have a low attendance record in CNSA meetings. The configuration of the CNSA forum may therefore partly explain the limited debate on non-agricultural aspects of food security and the emphasis put on the monitoring of farming seasons.

Nutrition and social dimensions of food security are discussed in other bodies: the National Coordination Council on Nutrition (Conseil National de Coordination en Nutrition, CNCN) chaired by the Ministry of Health (MoH) and the National Council on Social Protection (Conseil National sur la Protection Sociale, CNPS) recently created under the Prime Minister’s department. Therefore, there is no single institution covering all the dimensions of food and nutrition security,¹⁶ but rather a multiplication of inter-sectoral bodies dealing with each dimension, with the risk of overlaps and confusion. Non-agricultural actors prefer engage with institutions that directly address their respective fields. In particular, nutrition actors raise some concerns about the systematic combination of food security and nutrition, which could potentially lead to the merging of the CNSA and the CNCN. Nutrition is indeed determined by many factors (food, childcare, health and sanitary conditions) and these actors fear that there is a risk of it being reduced only to its food component and that specialist nutrition forums may consequently be delegitimised and left behind.

However, the rise of ideas such as “nutrition-sensitive interventions”¹⁷; or social transfers on the global political agenda is currently influencing national policy debates on food security through a reconfiguration of power games. In the wake of the country becoming a member of the Scaling Up Nutrition movement (SUN) in 2011 and of the World Bank’s action on social protection, the CNCN and the CNPS are increasingly claiming to address food security issues. One interviewee stated: “[on the issue of targeting the vulnerable] social protection increasingly enters food security and food security is more and more in social protection”. Nutrition actors benefit from the ability to circulate easily from one forum to another in Burkina Faso to advocate for a better integration of nutritional issues in food security policies in the CNSA. The insufficient integration of nutrition in food security policies is a recurring issue and is often linked to the institutional struggle between the MoA and the Ministry of Health. One interviewee in the nutrition field encapsulated well the issue: “when you ask a MoA representative to describe the state of food and nutrition security in Burkina Faso, you receive a good presentation of the methodology including data from nutritional surveys, but at the end only cereal production aspects are described. Nutrition is not the problem of the MoA which delegates the issue to the Ministry of Health whereas the MoA is also responsible (…) there should not be a war between Agriculture and Health”.

A few nutrition actors – who could be considered as “mediators” – regularly question members of the CNSA. They highlight the contradiction between a policy document dealing with food and nutrition security (PNSAN) and the existence of two different forums dealing respectively with food security and with nutrition. Historical agricultural actors are thus progressively obliged to widen their vision of food security to health and nutrition issues, just as they have had to do with environmental issues (Dury et al., 2015). For example, nutrition-sensitive agricultural interventions have consisted of the revision of the curricula of the national agriculture school to include nutrition courses in the basic training of agriculture extension agents (ACF, 2013).

5 Discussion

Insights from the combination of new institutional approaches and the cognitive analysis of public policy enable us to identify three main explanations for the locking in of food security policies around agricultural production in Burkina Faso. These explanations could constitute the core elements of a generic framework for institutional diagnostics of food security policies. Analysing food security policies through this institutional lens could allow the assessment of whether inter-sectorality in food security policies is worth promoting and testing. This also makes it possible to attune to signals of policy change and to see inter-sectorality as an evolving process.

The first core element of the institutional diagnostic refers to the predominant problematisation of food security supported by the majority and/or the more powerful actors involved in the field. It is essential to place the current conception of food security in a historical perspective to understand from where it comes, in what context it emerged and by whom it was supported. In the case of Burkina Faso, the tragic food crises of the 1970s have profoundly influenced the present-day problematisation of food security, which is obviously different from the situations in other countries. Institutional diagnostics of food security policies in a given country should therefore analyse the influence of history on the framing of food security, as path dependency may partly explain difficulties in policy change. Conversely, it is important to pay attention to current new ideas which constantly emerge, circulate and may challenge the dominant referential.

The second core element of the proposed framework for an institutional diagnostic corresponds to instruments built to characterise, monitor and evaluate the public policy problem. Apparently technical and neutral, these “micro-institutions” actually embed a particular idea of it. Consequently they can be strategically used by different groups of actors as a means to promote their vision. Analysing the food security information system in Burkina Faso helps us to understand the current focus of food security monitoring on food production data. This analysis also enables us to track pathways of change. The progressive integration in the SISA of data related to determinants of food security other than food availability undoubtedly influences the framing of food security debate and decision-making. Beyond the technical difficulties of integrating different information systems, discussions regarding prevention and management of food crises are now fuelled by an increasing diversity of information. Nutritional information, in particular, is now systematically presented in data monitoring farming seasons and often serves to nuance the food security situation suggested by only food production information.

Lastly, the institutional configuration of the policy debate and the nature of the different forums at stake are another core element to consider in institutional diagnostics. We have highlighted that the CNSA forum in Burkina Faso brings together representatives of the agricultural sector mainly, and thus it has had difficulties in making food security policies truly inter-sectoral, despite its inter-ministerial mandate. This situation is partly due to the historical legitimacy and the political resources available to agricultural actors, and to the fragmentation of the food security policy debate into different forums. Most actors do not challenge the situation which they consider as normal, as illustrated by the following statement from officials in communal health and social action services: “When we talk about food insecurity we immediately think of officials from agriculture. As health officials, we are in charge of nutritional issues”; “The term ‘food and nutrition security’ is not part of our daily vocabulary. We talk rather about food aid for the poorest of the poor”. Only a few policy actors try to fuel alternative visions in the food security policy debate.

The institutional configuration also influences on the way power games develop and may challenge actors’ positions, pushing or not for more inter-sectoral food security debates and policies. In the case of the CNSA in Burkina Faso, the voice of nutrition actors who support a vision of food security that is less directly linked to agricultural production and who convey global ideas such as nutrition-sensitive agriculture could influence the national debate. If they fully engage with this forum and develop strategies to increase their political resources, these actors could be agents of change (“mediators”) towards more inter-sectoral food security policies. The possibility of changing the power games among actors by enabling a broader policy debate is therefore to be considered in institutional diagnostics. Such an approach could be useful to understand why one particular idea such as inter-sectorality fails or succeeds in imposing and framing policy.

6 Conclusions

The recognition of the multidimensional nature of food security often leads to the idea that food security policies should be redirected towards inter-sectoral coordination. Although this idea makes sense, it is critically important to diagnose the way food security policies are shaped and implemented before blindly recommending inter-sectoral food security policies and institutional reforms. If not, the risk is that inter-sectorality is doomed to failure as was the case with multisectoral planning in nutrition.¹⁸

In this paper, we propose a framework for institutional diagnostics that enables understanding why food security policies tend to be more sectoral and unidimensional – around agricultural production – than inter-sectoral in Burkina Faso. In this West African country, though increase in agricultural productivity is certainly needed, the fact that food surplus areas experience higher levels of food insecurity and malnutrition than the national average shows that food insecurity is much more complex than simply a question of production. At the same time, we give insights on how to identify institutional signals (even weak ones) of policy changes.

Drawing on the combination of new institutional approaches and cognitive analysis of public policy, we explain the difficulties of making inter-sectorality a reality in food security policies in Burkina Faso. The history of food security was marked by tragic events in the 1970s and its problematisation around cereal deficits continues to have a strong impact on the way food security is framed today. Despite significant evolution, the food security information system continues to frame policy debate and decision-making around food production because administrative routines have been built on instruments monitoring crop seasons. Lastly, agricultural actors have a dominant position in food security forums, due to their historical legitimacy and to their higher political resources with respect to other actors.

The three main explanations – path dependency, measurement instruments and power games in policy forums – that explain lock-ins in food security policies in Burkina Faso form the core elements of the framework we propose for institutional diagnostics on food security policy. Such a framework can help policy-makers become aware of possible bias and lock-ins in food security policies by enlightening how predominant ideas are linked with specific measurement instruments and institutional configuration of policy debate. Conversely, such a framework can help them see that changes in measurement instruments of food security or power relations between actors in policy forums may be instrumental in driving policy change. The way nutrition issues have progressively entered food security information systems and the food security policy agenda can be seen as the premises of more nutrition-sensitive food security policies. Finally, in addition to be particularly relevant to capture institutional conditions at national level, our framework may also be useful at local levels in order to better understand and integrate policy implementation into the analysis. Indeed, decentralisation processes and bureaucratic traditions – which may vary from one area to the other – are undoubtedly crucial in the way food security policies are implemented on the ground and offer particularly interesting new avenues for research.

Notes

1 The evolution of the food security concept is a long history of enlargement from the initial focus on food availability to the importance of access to food with Sen’s works and other subsequent dimensions. Hence, a number of authors have highlighted the complexity of the issue, characterised by divergences, controversies and fragmentation in framings (see for instance Maxwell, 1996; Mooney and Hunt, 2009; Janin, 2010).

2 This is well illustrated in the discourse of José Graziano Da Silva (General Director of FAO): “It is crucial that countries adopt new modes of governance going beyond traditional ministries to which specific sectors are given in order to find innovative solutions to complex development problems” (Forum for the future of agriculture, 22 March 2016, Brussels). Inter-sectorality is also strongly advocated in nutrition (see Garrett and Natalicchio, 2011) and in its interactions with HIV/AIDS for instance (Hunsmann, 2010).

3 We concur with IEH (2012) which, based on the analysis of six country case studies, states: “Although Food Security and Nutrition (FSN) is conceptually understood in its multiple dimensions, and there is a wide consensus that FSN should be addressed inter-sectorally (i.e. FSN is an area in which different sectors need to work together), in practical terms the institutions that orient and finally execute the actions at the field level tend to be sectorally focused”.

4 Mediators can then be considered as “policy entrepreneurs” described by Kingdon (1984/1995) in explaining policy change.

5 Poverty is 59.4% (compared to 40.1% at the national level), food insecurity is 45.4% (compared to 35.4% at the national level) and acute malnutrition is 12.3% (compared to 10.4% at the national level and the WHO critical threshold of 10%). Regarding another western region of the country, Hauts-Bassins, Lourme-Ruiz et al. (2016) shows that relatively abundant cereal production (412 kg on average per capita per year over the period 2007–2011 compared to 261 kg on a national level and to 190 kg as the food self-sufficiency threshold) coexists with high prevalence of chronic malnutrition of children under 5 (33%).

6 The PNSAN has been adopted in a Ministers Council on October 2014 but no decree of application was drafted due to the social insurrection that toppled the President and led to a transition government.

7 “Given its transversal aspect, food and nutrition security requires that particular attention is paid to the strengthening of its governance (…) which will be done especially through an improved inter-sectoral collaboration” (: 38).

8 (1) Sustainably increase food availability to cover national needs; (2) Reinforce the capacity to prevent shocks and respond to shocks; (3) Improve physical and financial accessibility to food; (4) Improve the nutritional status of communities; and (5) Reinforce the governance for food and nutrition security.

9 The Ministry of Agriculture was renamed Ministry of Agriculture and Food Security (Ministère de l’Agriculture et de la Sécurité Alimentaire, MASA) in 2012, then changed under the transition government, in November 2014, to Ministry of Agriculture, Hydraulic Resources, Sanitation and Food Security (MARHASA), and yet again in November 2015 with the newly elected President to become the Ministry of Agriculture and Water Developments (MAAH).

10 The PNSAN will be operationalised through the second PNSR 2016–2020 which is currently being drafted.

11 All the Sahel countries experienced the same situation and this led to the creation in 1973 of the Inter-State Permanent Committee for the Fight of Drought in Sahel (Comité Permanent Inter-Etats de Lutte contre la Sécheresse dans le Sahel, CILSS) as the regional organisation in charge of preventing and managing food crises.

12 According to the PA-SISA “The evolution of the understanding and of the consideration of food security determinants other than food availability did not systematically lead to the necessary reforms of the statistical production and coordination system” (MAHRH, PA-SISA, 2004, p. 46).

13 The same observation has been made in Mali by Dury et al. (2010).

14 A parallel with the “seasonal” policy of social exclusion in France for instance can be helpful: it is indeed well known that the problem of homeless people only reaches the top of policy agendas in winter and this situation is systematically repeated every year (making this topic a seasonal news craze).

15 Bricas and Daviron (2008) and Fouilleux et al. (forthcoming) for example explain how the 2008 food crisis led to a return of “productionism” in global food security debates.

16 The situation in Burkina Faso is very different from the one in Brazil, for example, where the Inter-sectoral Food Security Council played a key role in addressing all dimensions of food security (Rocha, 2009; Leão and Maluf, 2012). This does not mean that a similar body should be recommended for Burkina Faso. Such a solution might not feasible, or more efficient, in the fight against food insecurity in a different context.

17 This refers to the series of articles on maternal and child undernutrition published in The Lancet in 2008 and in 2013, and the launch of the international movement Scaling up Nutrition SUN in 2010.

18 The review Food Policy hosted a famous debate between Field (1987) announcing the death of multisectoral nutrition planning (his article titled “Multisectoral nutrition planning: a post-mortem”) and Berg (1987) bringing a more nuanced response a few months later.