Diagnosing integrated food security strategies

Highlights

•	Many African governments have developed integrated food security strategies.
•	Little is known about the properties and outcomes of such integrated efforts.
•	The paper proposes three diagnostic steps for studying these.
•	The steps build on (i) policy integration, (ii) mechanisms and (iii) policy success.
•	The diagnostics are illustrated by applying them to South Africa’s Integrated Strategy.

Abstract

The global food price crises of 2007–8 and 2010 and subsequent policy debates have led to increased recognition that the drivers of food insecurity and associated policies transcend the boundaries of traditional governmental sectors and jurisdictions. Building on this insight, many governments of countries facing food insecurity have developed, or are in the progress of developing, integrated food security strategies. However, in spite of their recent popularity, to date little is known about the properties and outcomes of these strategies. This paper aims to help overcoming this gap by proposing a way of diagnosing the expected variety of integrated food security strategies and associated outcomes. Three diagnostic steps are put forward, each of which is linked to a specific theoretical perspective from the Public Policy literature. The first step concerns diagnosing the variety of IFSSs and is referred to as descriptive diagnostics. This type of diagnostics is suggested to be performed by using a policy integration perspective. The second step involves diagnosing what causes variety and change. This step is named explanatory diagnostics and revolves around what ‘mechanisms’ explain (dis)integration. The third diagnostic step focuses on diagnosing the outcomes of IFSSs and is referred to as evaluatory diagnostics. For this type of diagnostics a policy success and failure perspective is proposed. The applicability of these diagnostic steps and associated theories is illustrated through the case of South Africa’s Integrated Food Security Strategy. The paper ends with a discussion of promising methodological approaches and with raising some hypotheses and expectations about performing these types of diagnostics in a Sub-Saharan African context.

Keywords:

• Integrated food security strategy
• Food security
• Policy integration
• Governance
• Public policy

1 Introduction

Recent years have witnessed increased recognition of the crosscutting nature of food (in)security¹ among academics, policymakers, and other stakeholders. The world food price crises of 2007-8 and 2010 showed that states of food insecurity are affected by a broad range of determinants, including changing diets, yield gaps, effects of climate change, poor governance, social inequality, the functioning of the global trade system, biofuels production, and financial speculation (Headey and Fan, 2008; High Level Panel of Experts, 2011; Tadesse et al., 2014); many of which had been insufficiently appreciated in both the scholarly literature and the steering efforts of governments and international organizations. This range of factors makes clear that food insecurity, as a problem that policymakers need to engage with, transcends the boundaries of ‘traditional’ sectors, levels, and arenas of government (Misselhorn et al., 2012; Qureshi et al., 2015; von Braun and Birner, 2016).

Following on this recognition, food security commentators have propagated governance approaches that put greater emphasis on coordination, policy integration, or holism (Chen et al., 2016; Lang et al., 2009; Maxwell et al., 2010). The core assumption lying underneath such calls is that concerted efforts across jurisdictions will enhance the coherence and consistency of policy goals and instruments, eventually effectuating in better targeted food security interventions. For that purpose, many governments of countries facing severe food insecurity have recently developed, or are in the progress of developing, integrated food (and/or nutrition) security strategies (IFSSs), particularly also in Sub-Saharan Africa (Table 1). Integrated food security strategies are explicit governmental attempts to align or even fundamentally redesign goals, instruments, and capacities (cf. Rayner and Howlett, 2009), with the purpose of achieving the four food security dimensions of availability, access, utilization, and stability. As such, these strategies can be considered as ambitious and potentially far-reaching attempts to realize concerted action across governmental levels and sectors (e.g., agricultural, public health, trade, environmental, social and economic policy). Although many of these strategies have been developed in reaction to the food price crises and are thus fairly new, some older examples that have already attracted considerable scholarly interest exist. Two examples of the latter are Brazil’s ‘Fome Zero’ program (Hall, 2006; Rocha, 2009) and South Africa’s Integrated Food Security Strategy (Drimie and Ruysenaar, 2010; Hendriks, 2014; Pereira and Ruysenaar, 2012). Many of the more recent strategies in Sub-Saharan Africa (SSA) have been developed in close collaboration with international donors and think tanks, most notably the International Food Policy Research Institute (IFPRI), the Food and Agriculture Organization of the United Nations (FAO), the Scaling Up Nutrition movement (SUN), and the United States Agency for International Development (USAID).

Table 1 Overview of integrated food and/or nutrition security strategies in Sub-Saharan Africa.²

Country	Strategy	Years	Country	Strategy	Years
Angola	National Food Security and Nutrition Strategy	2009–2013	Madagascar	Plan National d’Action pour la Nutrition	2012–2015
Benin	Plan Stratégique de Développement de l’Alimentation et de la Nutrition	2009-?	Malawi	Food Security Action Plan	2008–2013
Botswana	?		Mali	Programme national de sécurité alimentaire; Politique Nationale de Nutrition	2006–2015; 2012–2021
Burkina Faso	Politique Nationale de Sécurité Alimentaire et Nutritionelle	2013-	Mauritania	Stratégie Nationale de Sécurité Alimentaire pour la Mauritanie; Plan d’Action Intersectoriel de Nutrition	2012–?; 2012–2015
Burundi	Programme national de sécurité alimentaire; Multi-sectoral Food and Nutritional Security Strategic Plan	2009–2015; 2014–2017	Mauritius	Food Security Plan	2008–2013
Cameroon	Politique et Programme Nationale d’Alimentation et de Nutrition	2007–2011	Mozambique	Multisectoral plan for chronic malnutrition reduction; Food Security and Nutrition Strategy	2011–2020; 2007–2012
Cape Verde	Estratégia Nacional de Segurança Alimentar e Nutricional	2015–2020	Namibia	National Food and Nutrition Policy; Strategic Plan for Nutrition	1995–?; 2012–2015
Central African Republic	Programme national d’investissement agricole, de sécurité alimentaire et nutritionelle; Politique Nationale de Nutrition de la République Centrafricaine	2013–2018; 2013–?	Niger	Plan National D’Action Pour La Nutrition	2007–2015
Chad	Programme National de Sécurité Alimentaire	2013–2020	Nigeria	National agriculture and food security strategy; National Strategic Plan of Action for Nutrition	2010–2020; 2014–2019
Comoros	?		Rwanda	National Food and Nutrition Policy	2013–2018
Congo	Cadre Stratégique de Lutte Contre la Malnutrition au Congo horizon 2025	2015–2025	São Tomé and Príncipe	Programme National de Sécurité Alimentaire et Nutritionelle	2012-
Côte d’Ivoire	Plan national de nutrition; Programme National de Sécurité Alimentaire et de Nutrition	2016–2020; 2009–?	Senegal	Plan de Réponse à l’Insécurité Alimentaire et Nutritionelle	2014-
Djibouti	Programme national de sécurité alimentaire	2012–2017	Seychelles	National Food Security Strategy	2012–2015 (?)
DR Congo	Programme National de Sécurité Alimentaire	2010-?	Sierra Leone	Sierra Leone National Food and Nutrition Security Policy	2012–2016
Equatorial Guinea	?		Somalia	Somali Nutrition Strategy	2011–2013
Eritrea	Eritrea Food Security Strategy; National Nutrition Strategy	2004-?; 2007–2011	South Africa	National Policy on Food and Nutrition Security for South Africa	2013–?
Ethiopia	National Nutrition Programme; Food Security Program	2013–2015;?	South Sudan	?
Gabon	Plan national d’investissement agricole, de sécurité alimentaire et nutritionelle	Already adopted?	Sudan	National Nutrition Policy and Key Strategies	2009–?
Gambia	National Food and Nutritional Security Programme	2011–2015	Swaziland	National Food and Nutrition Policy	2009 (adopted?)
Ghana	Naitonal Nutrition Policy	2014–2017	Tanzania	Tanzania Agriculture and Food Security Investment Plan	2011–2021
Guinea	Politique Nationale d’Alimentation − Nutrition	2005–2015	Togo	Programme National d’Investissement Agricole et la Sécurité Alimentaire; Plan Stratégique National pour l’Alimentation et la Nutrition	2011–2016; 2012–2016
Guinea-Bissau	Plano Estratégico de Nutrição	2015–2019	Uganda	Food and Nutrition Policy; Ugandan Nutrition Action Plan	2003–?; 2011–2016
Kenya	National Food and Nutrition Security Policy	2011–?	Zambia	National Food and Nutrition Strategic Plan for Zambia	2011–2015
Lesotho	National Action Plan for Food Security	2007–2017	Zimbabwe	National Nutrition Strategy; Food and Nutrition Security Policy in the Context of Economic Growth and Development	2014–2018; 2012–? (adopted?)
Liberia	National Food Security and Nutrition Strategy	2008-?

Given the involvement of these various international actors, the different food security contexts countries face, and differences in national governance systems and traditions, one would expect variation in the design as well as outcomes of IFSSs. However, in spite of their increasing popularity and associated high expectations, to date only few scholarly attempts to study these strategies and their outcomes have been undertaken. In addition, of the studies that have been conducted, various restrict their analysis to changes in eventual food security or nutrition indicators after an integrated strategy has been adopted (Agaba et al., 2016; Gubert et al., 2017) without elucidating whether the correlations they find result from any substantial governance changes. Others have opened up the black box of what happens within governance systems (e.g., Harris et al., 2017; Kampman et al., 2017; Pomeroy-Stevens et al., 2016), providing valuable insights into the difficulties of (and some successes in) translating paper commitments into effective and genuine governance changes. At the same time, these accounts have remained limited to individual cases and/or generally do not engage with current theoretical debates within the fields of Public Administration, Governance, and Policy. As a result, they offer few lynchpins for moving towards a more systematic evidence base and theoretical synthesis (cf. Purdon, 2014). This is not only problematic from a scholarly perspective, but also hampers well-informed policy recommendations (cf. Rodrik, 2010).

This paper contributes to the Special Issue’s interest in performing institutional diagnostics for food security in Sub-Saharan Africa by showing how three theoretical perspectives that have recently emerged in the field of Public Policy can be used to further research on IFSSs. Adopting these perspectives in future research would allow for realizing a coherent comparative research agenda. The three perspectives correspond with three types of diagnostics that appear particularly pressing for research on IFSSs: i) diagnosing the variety of IFSSs; ii) diagnosing what causes variety and change; and iii) diagnosing the outcomes of IFSSs. The first diagnostic step is here referred to as descriptive diagnostics and will be argued to be best studied through a ‘policy integration’ perspective. The second step is named explanatory diagnostics and revolves around what ‘mechanisms’ explain (dis)integration. The third type involves evaluatory diagnostics and deals with how to determine IFSSs’ success or failure. Whereas these three diagnostic steps are more concerned with policy than institutional variables, institutions will be shown to play an important role. In addition, although the three diagnostics are specifically linked to emerging questions about IFSSs in this paper, they could in principle also be used to study other types of (food) policy interventions.

The following sections elaborate each of the three diagnostics and associated theoretical perspectives. Subsequently, the applicability of the diagnostic steps is illustrated through the case of South Africa’s Integrated Food Security Strategy, which is probably the most intensively studied IFSS in Sub-Saharan Africa. For this secondary analysis I draw on a body of existing studies and accounts (Drimie and Pereira, 2016; Drimie and Ruysenaar, 2010; Hendriks, 2014; Pereira and Drimie, 2016; Pereira and Ruysenaar, 2012). The aim of these insights is purely illustrative; a more comprehensive analysis of South-African food security governance can be found in the contribution of Termeer et al. (XXX) in this Special Issue. The paper ends with a discussion of promising methodological approaches and with raising some hypotheses and expectations about performing these types of diagnostics in a Sub-Saharan African context.

2 Descriptive diagnostics: IFSSs as part of policy integration processes

A first diagnostic step entails describing IFSSs’ main properties and consequential variety. Although these strategies, as ambitious and often publicly welcomed programs or visions, naturally stand out on themselves, they can be considered as manifestations of broader attempts at strengthening policy integration vis-à-vis food security. Policy integration and similar concepts, like coordination, joined-up policy, or holistic governance, have been subject to various conceptualizations within the public policy literature (for detailed discussions, see: Cejudo and Michel, 2017; Peters, 2015). Here, it is understood as an agency-driven (see Section 3) and multi-faceted process that shapes the degree to which a polity and associated policies address a crosscutting policy problem, here food insecurity, in a more or less holistic manner (Candel and Biesbroek, 2016). The merit of approaching policy integration as a process (as opposed to, for example, a desired principle or outcome) and as multi-faceted (i.e. consisting of various dimensions) is that it allows for analyzing variety as well as change over time.

The four dimensions that constitute policy integration processes and that would thus be central to descriptive diagnostics are: (i) policy frames, (ii) subsystem involvement, (iii) policy goals, (iv) and policy instruments (Candel and Biesbroek, 2016). Whereas the first two are primarily related to institutional variables, the latter two are focused on the policy level. Crucially, these dimensions often differ in the extent to which they are integrated. For example, sustainable development scholars have shown that many governments have adopted comprehensive sustainable development strategies (i.e. have well-integrated goals), but have lagged behind in terms of designing integrative mixes of policy instruments (Jacob et al., 2008). Thus, the four dimensions may move at a different pace, or even in opposite directions, i.e. when disintegration on one or multiple dimensions occurs. Table 2 provides an overview of what the four dimensions may look like for different degrees of policy integration in food security governance. Importantly, these are ideal-type manifestations; they can be used as points of reference for descriptive diagnostics, but, depending on the specific context, considerable variation is possible.

Table 2 A policy integration framework for descriptive diagnostics, adopted from Candel and Biesbroek (2016).

	Lack of policy integration			Full policy integration
Policy frame	Food security is defined in narrow terms, for example as solely an issue of agricultural production. Other types of efforts, e.g., health promotion or social policy, are not understood in terms of food security.	Realization that (externalities of) policies of various domains influence food security outcomes. Food security defined in narrow terms, but some attention to other dimensions and interactions with other policy efforts.	Awareness that food (in)security is affected by a broad range of factors and influences. Food security is viewed as a multi-dimensional policy problem that spans the boundaries of various policy domains, which are therefore believed to be needing to coordinate their policy efforts.	Recognition that achieving food security, across all its levels and dimensions, is a challenge to the governance system as a whole. Roles of subsystems are viewed as part of a holistic approach. Integration is advocated and adopted as dominant steering philosophy.
Subsystem involvement:	Food security falls within jurisdiction of a dominant subsystem, e.g., agriculture. Formally no other subsystems are involved, although they may an effect on food security.	Idem, but food security concerns emerge in one or more additional subsystems.	Food security (in a broader sense) formally embedded within more subsystems, e.g., trade and agriculture.	Food security is embedded within all potentially relevant subsystems. Subsystems that are formally involved is equal to or higher than at previous manifestation, but complemented with loosely coupled set of alternative subsystems.
Subsystems involved
Density of interactions	No interactions	Infrequent information exchange between dominant subsystem and one or more alternative subsystems.	More regular and formal exchange of information and coordination, possibly through coordinative instruments at system-level.	High level of interactions between formally involved subsystems, that maintain infrequent interactions with loosely coupled set of subsystems.
Policy goals:	Sector-specific food security goals in one subsystem. For example: food security as priority in agricultural policy.	Idem, but in addition food security concerns may be embedded to some extent in other subsystems’ policies.	Various subsystems adopt food security concerns in their policy goals, some subsystems may develop comprehensive sectoral food security strategies.	All potentially relevant subsystems have adopted food security concerns in their policy goals.
Range of policies in which problem is embedded
Policy coherence	Very low or no coherence.	Externalities of some policies may be addressed to some extent.	Coordinated sectoral goals, which are judged in the light of coherence. Subsystems attempt to develop synergies.	Shared policy goals embedded within an overarching integrated food security strategy.
Policy instruments:	Sector-specific food security instruments in one subsystem. For example: food aid or agricultural subsidies.	One or more additional subsystems (partially) adapt their instruments to address food security concerns or mitigate negative effects.	Various sectoral instruments and instrument mixes that are intended to contribute to food security.	All potentially relevant subsystems have adopted food security instruments in their policies.
Range of subsystems’ policies that contain policy instruments
Procedural instruments at system-level Consistency	No relevant procedural instruments at system-level.	Some information sharing instruments, such as inter-service consultations or impact assessments, may be initiated.	Range of cross-sectoral instruments, such as impact assessments or coherence mechanisms.	Broad range of procedural instruments at system-level, including boundary-spanning structures that coordinate, steer and monitor subsystems’ efforts. E.g., presidential/prime-minister desk for food security
Consistency	Very low or no consistency.	Some obvious externalities, e.g., of agricultural or trade policies, may be reduced.	Attempts to reduce all externalities and to realize synergies and positive food security spin-offs.	Full reconsideration of subsystem instrument mixes, resulting in a comprehensive, cross-subsystem instrument mix that is designed to meet a set of coherent goals.

The dimension of policy frame concentrates on how the problem of food security is perceived within a government or governance arrangement and includes the institutionalized norms and beliefs. Is food security, for example, predominantly framed as primarily a matter of increasing agricultural production, or are social-economic, environmental, and health issues taken into account as well? The central question for this dimension is whether the crosscutting nature of food security is recognized as such, and whether the need for an integrative approach is acknowledged (cf. Peters, 2005). Of course, different frames may exist within a polity. For example, a department of health will approach food security from a different perspective than a department of agriculture (cf. Candel et al., 2014). However, some frames may find wider resonance than others − particularly also amongst high-level decision-makers − and as such have a greater impact on the direction of policy interventions (Baumgartner and Jones, 2009; Lau and Schlesinger, 2005).

Subsystem involvement concerns the range of actors and institutions that are involved in governing food security within a polity. Subsystems are relatively closed and stable actor-networks that discuss and shape policies in a certain domain or vis-à-vis a certain topic (Zafonte and Sabatier, 1998). These networks can be discerned within governments, but also as parts of broader governance constellations in which both public and private actors are active. For example, the roles of international organizations, donors, and NGOs would be highly relevant in the context of Sub-Saharan African food security governance (Devereux, 2009). As food security potentially crosscuts many domains and topics (Misselhorn et al., 2012), the central question here is which of these subsystems are (or are not) involved. In addition, this dimension is about the density of interactions between subsystems. Some subsystems may be involved to a larger extent than others, as some policies are (perceived as) more relevant to food security dimensions than others. For relatively higher degrees of policy integration one would expect food security to be primarily embedded within a limited number of subsystems, which maintain frequent interactions with a set of loosely coupled subsystems that are engaged to a lesser extent (cf. Orton and Weick, 1990).

Third, the dimension of policy goals is about the range of policies in which food security concerns have been explicitly adopted as well as the coherence between these goals. Ideally (but far from always), the involvement of a broader range of subsystems would translate into a similar broadening up of subsystem policies in which food security concerns are being addressed. Integrated food security strategies fall under this dimension; they can be considered as the most integrated form of goals (cf. Metcalfe, 1994). Overarching strategies are used to explicitly set out how the policy efforts of individual subsystems may collectively contribute to shared (food security) goals (Rayner and Howlett, 2009), thereby ensuring their coherence or even facilitating synergies. As the operationalization and measurement of such coherence have not (yet) been agreed upon (Nilsson et al., 2012), this remains largely to the interpretation of the analyst.

The fourth dimension is constituted by the policy instruments that are (or are not) deployed to pursue food security goals. These instruments can be embedded within an overarching strategy, but do not need to be. Under this dimension, three aspects are of relevance. First, in relatively more integrated food security approaches, in which a broader range of subsystems are involved and policy goals are adopted, there would ideally be a diversification of policy instruments, both in types and across sectors. Public policy scholars have devoted considerable effort to developing instrument typologies (see Hood (2007) for a review of the policy instrument literature. See also Hood (1983); Howlett (2005, 2009); Lascoumes and Le Gales (2007); Verdung (1998) for different examples of instrument typologies). Interestingly, whereas the literature on environmental governance has come up with a well-developed typology of specific environmental policy instruments (Jordan et al., 2005; Wurzel et al., 2013), such an attempt has not yet been undertaken for food (security) governance. Hood’s (1983) classic typology of instruments drawing on the governing resources of nodality (information), authority, treasure (financial resources), and organization seems a particularly suitable point of departure for comparing instrument mixes and associated calibrations (see Daugbjerg and Sønderskov, 2012 for an elaborate discussion of distinguishing relatively ‘harder' and ‘softer' calibrations). Second, integrative approaches can be supported by the deployment of procedural instruments, i.e. policy instruments that are aimed at altering the functioning of a polity itself (Howlett, 2000), in order to facilitate subsystem interactions and coordination. Examples of such procedural instruments are inter-departmental task forces or constitutional provisions. Third, this dimension is about the consistency of the entire policy instrument mix, which is relative to the (more or less) coherent set of policy goals (Howlett and Rayner, 2007). An appropriate instrument mix effectively contributes to pursuing these goals (cf. Cirillo et al., 2017; Qureshi et al., 2015). However, as with the coherence of goals, there is no agreed-upon method for assessing the consistency of instruments, thus putting much weight on the researcher’s or experts’ interpretation (Nilsson et al., 2012).

Using the framework for performing descriptive diagnostics can contribute to addressing a number of pressing questions about the emergence of IFSSs. A first question is whether the emergence of integrated food security strategies, i.e. strengthened policy integration in terms of goals, has been accompanied by shifts in the other dimensions. This question is particularly pressing because research on the governance of other crosscutting problems has shown that many integrative strategies do not proceed beyond paper realities (Candel 2017; Jacob et al., 2008; Jordan and Lenschow, 2010; Mickwitz and Kivimaa, 2007). Second, the framework allows for comparing policy integration in different phases of the policy cycle (Howlett and Ramesh, 2003). For example, goals and instruments may be well integrated in policy design, but less so in the implementation (cf. Drimie and Ruysenaar, 2010; Mosse, 2004), or vice-versa. Third, it enables comparing differences in food security governance across countries as well as over time. Such comparisons are essential for obtaining a better understanding of food security governance, which would allow for posing follow-up questions about what explains variation of policy integration processes and effects (the next diagnostic step).

3 Explanatory diagnostics: a mechanismic perspective

Although the abovementioned policy integration framework helps diagnosing shifts in degrees of policy integration (dimensions) over time, it does allow moving beyond description. Therefore, a second diagnostic step entails complementing the framework by uncovering the processes that explain why policy (dis)integration occurs or, more specifically, what explains the emergence of IFSSs as well as their variation and change over time. Even though scholars have identified long lists of factors affecting policy integration (for an overview, see: Peters, 2015), many of these factors have proven rather unsatisfactorily, as they are rather static and provide limited insights into the social processes through which they are mediated (Biesbroek, 2014; Biesbroek et al., 2015). Perhaps most fundamentally, various factors have been shown to have the potential of both hindering and contributing to policy integration, depending on the specific institutional context and governance interactions in which they occur (Peters, 2015). For that reason, scholars have advocated a shift towards studying the causal mechanisms driving processes of policy integration (Biesbroek and Candel, 2016; Biesbroek et al., 2014).

Causal mechanisms are here understood as interactions between actors (e.g., decision-makers, civil servants, donors, interest representatives) that produce specific outcomes, and as such mediate between dependent (the policy integration dimensions) and independent variables (Hedström and Ylikoski, 2010). These interactions are believed to occur in certain patterns, the analysis of which can result to identifying abstract categories of mechanisms. The specific functioning of a mechanism and the outcome(s) it produces are determined by the characteristics of the actors (e.g. ideas, norms, values) and the (institutional) context in which the mechanism is triggered (Biesbroek et al., 2014). As such, mechanisms are about the dimension of politics, which is affected by but also adjusts and reproduces the polity and policy contexts. Put differently, this diagnostic step brings in the role of agency in accounting for why and how policy integration in food security governance shifts over time. Previous studies have shown that, more than institutional structures or policy frames, it is these type of actor interactions that explain why many attempts at strengthening policy integration get stranded (Peters, 2015; Vince, 2015).

A challenge to using mechanisms as theoretical perspective is the relative newness of this branch of research, particularly in the context of policy integration studies. It is therefore impossible to provide a comprehensive list of mechanisms that could explain policy integration outcomes in food security governance beforehand. Instead, these mechanisms will have to be distilled abductively through process-tracing approaches that pay careful attention to both the behavioural patterns as well as the institutional contexts within which these occur (cf. Beach and Pedersen, 2013; Section 6.1). Table 3 provides an overview of some indicative mechanisms that have been found in previous studies. Although these mechanisms show a large variation in scope and some are very generic, they provide an indication of the type of interaction patterns that are considered relevant. It should be noted that the causal mechanisms that are deemed relevant here are not necessarily policy integration mechanisms; they may also occur in and contribute to other types of institutional and policy change processes.

Table 3 Examples of causal mechanisms driving policy integration processes.

Jochim and May (2010)^3	Biesbroek et al. (2014)^4	Biesbroek and Candel (2016)
Social learning: interactions through which actors within and between subsystems learn about the (crosscutting) nature of the problem and its governance.	Conflict infection: transposition of conflicts in one policy arena to other arenas by the actors that move between these arenas.	Scientific advocacy: scientific actors pushing policymakers for developing better integrated policy.
Coalition building: alignment of powers within and between subsystems, possibly allowing for more integrative approaches.	Risk innovation mechanism: governments communicate about public innovations through risk minimizing strategies, while citizens want to negotiate about which levels of risks are acceptable.	Making one’s mark: decision-makers wanting to make their mark when being in office by developing ambitious (integrative) policies.
Policy entrepreneurship: the coupling of more integrative approaches to a perceived lack of integration during a perceived window of opportunity.	Frame polarization: interactions between actors in which the distance between opposing groups increases due to repeated reaffirmation of values and beliefs, without accepting the other’s point of view.	Inter-governmental learning: learning across governmental levels and agencies about experiences with integrated policy development.
		Inter-departmental competition and turf wars: struggles between departments over jurisdictions and resources.
		EU compliance: actors pushing for compliance with (integrative) EU legislation.

4 Evaluatory diagnostics: programmatic and political evaluation

A third diagnostic step involves the evaluation of the outcomes of integrated food security strategies, and is referred to as evaluatory diagnostics. Recent debates on policy success and failure seem to provide a useful lens for performing such diagnostics. The success and particularly the failure of public policy programs have been an important topic throughout the development of Public Policy as a scientific discipline. At the same time, conceptualizations and accounts of policy performance remained relatively fragmented for a long time. A revival of interest to questions surrounding policy success and failure in recent years has overcome much of this fragmentation and resulted in increasing conceptual clarity and convergence (e.g., Bovens and ‘t Hart, 2016;Howlett et al., 2015; Marsh and McConnell, 2010; Zittoun, 2015). It goes beyond the scope of this paper to synthesize this whole literature. Instead, I restrict myself to elaborating two different perspectives on evaluation that have been put forward by Mark Bovens and Paul ‘t Hart (Bovens, 2010; Bovens and ‘t Hart, 1996,2016): programmatic and political evaluation. Food security governance analysts can use these perspectives in two ways (see below): i) to directly evaluate the outcomes of integrated food security efforts, or ii) to study how other actors evaluate or make sense of these efforts.

Programmatic and political evaluation can be considered two different ontological perspectives, drawing on different strands of literature, to make sense of how policy outcomes are appraised. Broadly speaking, whereas programmatic evaluation is primarily concerned about a policy’s performance, political evaluation focuses on its reputation (Bovens and ‘t Hart, 2016). A programmatic evaluation rationale takes a rational choice point of departure; by studying observable facts, costs and benefits, and by comparing outcomes with original intentions, an evaluator can make claims about a policy’s effectiveness and efficiency in achieving its initial goals and addressing a problem. Programmatic evaluation revolves around ‘performance as measured’ and claims to be ‘objective’ or ‘evidence-based’ (Bovens & ‘t Hart, 2016, p. 4). The policy integration framework introduced in Section 2 could be used as a tool for performing such an evaluation. Political evaluation, on the other hand, approaches success and failure as being socially constructed in interactions between actors; thus taking a social constructivist point of departure. Policies and their performance are subject to political debate and contestation, because of which subjective claims rather than objective evidence become the prime focus. Such claims take the form of narratives and frames in which various problems, solutions, heroes and villains are brought to the stage. In addition, these social construction processes do not only occur in formal policy arenas, but also in the wider public domain. Political evaluation is thus about ‘reputation as conferred’ (Bovens and ‘t Hart, 2016).

Applied to integrated food security efforts, a first option would be that the analyst him- or herself performs an evaluation following one of the two perspectives. In a programmatic evaluation, the analyst could compare the goals laid out in integrated food security strategies with (changes in) food security indicators and, possibly, additional evidence about costs, externalities, and alternatives, to arrive at a balanced judgement of a strategy’s effectiveness and efficiency (cf. Barrett, 2010). A political evaluation would take a more normative point of departure; the analyst could, for example, study to what extent ‘good governance’ principles have been adhered to in food security governance or uncover ignored perspectives (e.g., McKeon, 2015).

A second and perhaps more interesting option from the perspective of institutional diagnostics would be that the analyst studies how others evaluate the outcomes of integrated food security efforts, and what institutional dynamics (e.g., rules, beliefs, arenas) occur and prevail. Questions that could serve as point of departure include: how are food security efforts evaluated in official studies, reports, and audits? What assumptions, logics, and procedures were followed? How do different coalitions of actors construct story lines about food security efforts? Which ideas and interests lie behind these views? Which story lines prevail and why?

Probably unsurprisingly, the first types of questions (i.e. programmatic evaluation) are much more dominant in policy and scholarly debates about food security (cf. Candel, 2014; Delaney et al., 2016). This type of evaluation allows analysts to make judgments about advancements in ‘actual’ states of food (in)security, which is the ultimate goal or mandate of many government programs and international organizations. At the same time, the lack of attention to the social construction of food security strategies’ success and failure is problematic, as it has resulted in a lack of understanding of how ultimate beneficiaries as well as other public and private actors experience and judge food security efforts (cf. Yanow, 1996). Such an understanding would be crucial for fitting food security interventions to individuals’ (Bovens et al., 2001) (perceived) needs. In addition, it might serve as a basis for developing policy frames that find wide(r) resonance and may therefore allow for more sustained food security commitments. Because programmatic and political success are not necessarily correlated (Bovens and ‘t Hart, 2001), it is important to draw lessons following from both types of evaluation. Therefore, the central argument here is that both types of evaluation can produce complementary insights.

5 Illustrating the perspectives: South Africa’s Integrated Food Security Strategy

This section demonstrates the types of insights that can be acquired by applying the three diagnostic steps. For this purpose, I use the case of South Africa’s Integrated Food Security Strategy (2002-13), which is probably the most profoundly documented and analyzed IFSS in Sub-Saharan Africa. The section draws on previous analyses of South African food security scholars (Drimie and Pereira, 2016; Drimie and Ruysenaar, 2010; Hendriks, 2014; Pereira and Drimie, 2016; Pereira and Ruysenaar, 2012); no new data was collected. Importantly, the aim of this section is not to provide a comprehensive assessment of South Africa’s food security efforts (see Termeer et al. in this special issue), but merely to illustrate the applicability of the three perspectives for performing diagnostics.

The Integrated Food Security Strategy was developed in 2002, at a time of rising food prices, to combine a wide range of existing but fragmented food security efforts into an overarching approach. On paper, the strategy was embedded across a range of relevant sectors and levels of government and encompassed goals and instruments that targeted the various dimensions of food security. Nonetheless, the strategy was not the first IFSS to come into existence in South Africa; it closely resembled the Food Security and Nutrition Strategy adopted ten years earlier, much of which had remained unimplemented (Pereira and Ruysenaar, 2012). As will become clear, the IFSS suffered from similar implementation gaps and was replaced by a new Food and Nutrition Security policy (FNS) in 2013.

5.1 Descriptive diagnostics

The above shows that the development of the IFSS was a clear example of an explicit attempt at strengthened policy integration in the South African governance of food security. Applying the first diagnostic step, i.e. the processual policy integration perspective, can help to examine whether and how these changes actually occurred. The design and adoption of the IFSS show a shift from relatively ‘siloed’ goals and instruments into a more coherent and consistent set of goals and instruments in the strategy’s design (Drimie and Ruysenaar, 2010; Pereira and Ruysenaar, 2012). In addition, the creation of various coordinative instruments was envisioned, such as national and provincial coordinating units, local food security officers, and a National Food Security Forum (NFSF) for facilitating consultations with private stakeholders (ibid.). Although the National Department of Agriculture (NDA) was put in the lead of (coordinating) the strategy, other subsystems of public and private actors were given an explicit role as well. Furthermore, calling the new approach an integrated strategy could be seen as an attempt at creating an integrative policy frame across government. All in all, these steps suggest an overall shift towards strengthened policy integration.

Turning to the implementation of the IFSS, however, a very different image arises. Studies of the IFSS almost unanimously conclude that the strategy was implemented in a far for complete manner (Drimie and Ruysenaar, 2010; Hendriks, 2014; Pereira and Ruysenaar, 2012). The initially broad range of goals and instruments was restricted to a predominant focus on production increases, which can be explained by NDA’s dominant role (Drimie and Ruysenaar, 2010). Other subsystems participated infrequently or, in the case of provincial levels, used their autonomy to refrain from active implementation (Drimie and Ruysenaar, 2010). Procedural instruments, such as the NFSF or the Integrated Food Security and Nutrition Task Team, an integrated planning platform, were not implemented as originally intended. In addition, a shared problem definition never developed. Instead, stakeholders had very different understandings of food security policy and associated causes and effects (Hendriks, 2014; Pereira and Drimie, 2016), or of what a policy is to begin with (Hendriks, 2014).

Table 4 schematically depicts these observations by linking them to the policy integration framework (Table 2). The table shows how the framework can be used to assess changes in policy output over time as well as whether discursive forms of policy integration (i.e. on paper) result in actual changes of governance. The latter proved hardly the case; broad goals were narrowed down and departments continued functioning in a siloed and non-coordinated manner. As a result, analysts have drawn hard conclusions, which is where this diagnostic step goes over in evaluatory diagnostics. Drimie and Ruysenaar (2010, p. 324), for example, conclude that “these arrangements [the institutional arrangements to implement the IFSS] are largely superficial and overridden by departmental line-function demands on budgets and personnel, rather than providing true integration … the ‘institutional architecture’ of the IFSS remains no more than scaffolding.”

Table 4 Initial policy integration assessment of South-Africa’s IFSS.

	Functioning of Food Security and Nutrition Strategy (1992- 2002; Pereira and Ruysenaar 2012)	Official IFSS (2002)	Actual functioning IFSS (2002-13)
Policy frame	Official recognition (in strategies and governmental committees) that food security is a crosscutting problem that needs to be governed holistically. However, policy frame is never truly institutionalized across government; instead, constantly fades away and reappears.	Recognition that achieving food security, across all its levels and dimensions, is a challenge to the SA governance system as a whole. Roles of subsystems are viewed as part of a holistic approach. Integration is advocated and adopted as dominant steering philosophy.	Food security predominantly framed in terms of production. Lack of a shared problem understanding; only some attention to other dimensions and interactions with other policy efforts.
Subsystems involvement	Food security falls within jurisdiction of NDA. Some, but limited, involvement of and interactions with other departments.	Food security formally embedded within broad range of departments, although leading role for NDA.	Food security falls within jurisdiction of NDA. Some, but limited, involvement of and interactions with other departments.
Policy goals	Official holistic goals quickly faded away. Food security concerns primarily adopted in NDA policy.	Shared policy goals embedded within overarching IFSS. Not yet embedded in individual subsystems’ policies.	Shared goals hardly adopted in goals of individual subsystems, except NDA policies.
Policy instruments	Range of instruments, but not streamlined and considered ineffectual. Number of coordinative procedural instruments throughout the years (e.g., committee of experts and working group).	Consistent range of instruments targeted at addressing food insecurity. Number of coordinative procedural instruments envisioned.	Inconsistent instrument mix, emphasis on instruments promoting agricultural production. Procedural instruments hardly implemented.

This initial diagnostic assessment could easily be extended. It would, for example, be interesting to analyze whether the recent FNS has been more successful in realizing a change of governance. In addition, what would be particularly relevant in light of this paper and special issue is to compare these policy integration dynamics with those in other Sub-Saharan African countries (e.g., Harris et al., 2017; Kennedy et al., 2016).

5.2 Explanatory diagnostics

Turning to explaining the observations obtained by applying the policy integration framework, research on South African food security governance has provided a long list of factors that impeded a genuine shift towards higher degrees of policy integration after the adoption of the IFSS. Contrarily, hardly any (potentially) enabling conditions have been identified. Factors that were considered particularly constraining include the lack of capacity and resources, the NDA’s dominant role, organizational cultures, turf wars, existing ‘rules of the game’, the autonomy of lower governments, poor design, and a lack of high-level political leadership (Drimie and Ruysenaar, 2010; Hendriks, 2014; Pereira and Drimie, 2016; Pereira and Ruysenaar, 2012).

Although most of these static institutional factors are independent variables (explaining a lack of implementation) rather than intervening causal mechanisms, many can serve as point of departure for distilling such mechanisms. The rationale behind such an undertaking is that without studying how they play out in actor-interactions (in a particular context), these factors remain relative ‘black boxes’ in terms of their explanatory value (cf. Biesbroek et al., 2015; see Section 3). ‘Turf wars’ is an example of a causal mechanism, albeit a relatively generic one, that intervenes between variables such as interests, ideas, and jurisdictions on the one hand and the dependent variable of non-implementation on the other. At a lower level of analysis, this mechanism may be broken down in a more varied range of ‘sub-mechanisms’, such as departments attacking each other’s proposals or departments mobilizing their clients to prevent a reallocation of jurisdictions (, pp. 32−33).

It is difficult, if not impossible, to determine the exact mechanisms for South Africa’s IFSS, how these mechanisms interact, and how they relate to their specific contexts, without first performing an in-depth process-tracing analysis (Beach and Pedersen, 2013). Doing such an analysis would therefore be an important next step. A good example of a constraining interaction pattern that proved decisive for the IFSS’ non-implementation is what Drimie and Ruysenaar (2010, p. 329) refer to ‘juniorisation’; because ministers and directors-general hardly showed up at meetings and delegated junior staff, important decisions about could not or insufficiently be made.

5.3 Evaluatory diagnostics

Zooming in on how South Africa’s IFSS has been evaluated, a general observation is that the literature synthesized in this section has predominantly used a programmatic evaluation perspective.⁵ Many researchers do not necessarily measure the strategy’s performance in a strict sense, but do use some sort of implicit or explicit analytic criteria to assess the strategy’s outcomes and impacts. Comparing the strategy’s outcomes with its initial aims, most researchers conclude that the IFSS failed to deliver on its promises (Drimie and Ruysenaar, 2010; Hendriks, 2014; Pereira and Drimie, 2016; Pereira and Ruysenaar, 2012). This conclusion is drawn both in relation to the functioning of South African food security governance as with respect to eventual food system outcomes. Furthermore, it was observed that apart from these scholarly evaluations, few other explicit attempts at reviewing or evaluating the IFSS’s outcomes were undertaken; the government, for example, did not put up a monitoring and evaluation framework (Drimie and Ruysenaar, 2010; Hendriks, 2014).

From a political evaluation perspective, it seems as if there has not been an extensive public debate about the IFSS’s success or failure.⁶ It may also simply be the case that scholars have paid little attention to existing socio-political contestations regarding the strategy’s outcomes, which would then be an important avenue for future research (see Section 4). The only reflection of this type I found was a mentioning of NDA officials themselves reporting the implementation of a part of the IFSS (the Food Emergency Scheme) as a hasty work, not meeting the worsening food insecurity situation (Drimie and Ruysenaar, 2010). Of course, the accounts of researchers themselves also play an important role in the social construction of success or failure. The abovementioned (negative) evaluations and associated conclusions can be considered authoritative contributions to the debate (e.g., see: Ruysenaar, 2008), which are at least potentially influential in shaping public perceptions.

Furthering this type of diagnostics in the case of South Africa’s IFSS would require the analyst to be (more) receptive to how political actors and societal stakeholders make sense of the strategy’s functioning as well as how underlying institutional beliefs, norms, and implicit and explicit evaluation rules and procedures affect these evaluations. As argued above, the perceptions of ultimate beneficiaries are hereby not to be overlooked.

6 Discussion

This paper has proposed and illustrated three theoretical perspectives from the field of Public Policy that can be used for comparatively studying emerging integrated food security strategies along three diagnostic steps. To further research on integrated food security strategies a first step would, quite naturally, be the application of the perspectives to concrete empirical cases. As Section 1 showed, a majority of SSA governments have adopted some form of food and/or nutrition security strategy. However, the sparse evidence that does exist gives rise to serious doubts about whether these strategies are fitting, well-implemented, or, most importantly, (perceived as) effective on the ground (Drimie and Ruysenaar, 2010; Pereira and Ruysenaar, 2012; Sahley et al., 2005), although some new studies also cautiously observe small improvements (Harris et al., 2017; Kampman et al., 2017). At the same time, current understandings of the functioning of SSA public administrations are restricted by a shortage of descriptive research (Hyden, 2010), underlining the need for expanding the evidence base of African public governance in general, and food governance specifically. In this last section, I briefly reflect on some methodological approaches that prove promising for empirical inquiries. In addition, some expectations are raised about applying the perspectives to a Sub-Saharan African context.

Before doing so, however, it should be stressed that the diagnostic steps put forward in this paper have their limitations, i.e. there are boundaries to what they can be used for. Most importantly, the perspectives’ rootedness in the Policy Sciences makes that the steps are by definition more suitable for studying variety and change of policy variables than for accounting for the role of polity (i.e. the institutional context) and politics in food security governance. Although the paper has at various points pointed at the interactions between policy, polity and politics, and the mechanismic perspective takes political interactions explicitly into account, a better understanding of these two spheres would require bringing in additional theories. The other contributions in this Special Issue provide various fruitful suggestions for doing so. Secondly, although the third diagnostic step engages with IFSSs’ outcomes, determining policies’ outcomes or impacts in a programmatic sense remains a weak point of the Policy Sciences (or a step still to be taken). Other scientific disciplines, such as Economics and Public Health studies, may, despite facing challenges of their own, provide more fruitful methodologies for such assessments (Beekman et al., 2014; Sayinzoga et al., 2016).

6.1 Methodological approaches

A process-tracing methodology (Beach and Pedersen, 2013) seems most appropriate for performing the descriptive and explanatory diagnostic steps. Process-tracing is the systematic examination of, primarily qualitative, evidence to draw descriptive and causal inferences, and is mostly used to obtain a better understanding of a temporal sequence of events or phenomena (Collier, 2011). For descriptive diagnostics, process tracing methods can be used to systematically reconstruct how food security policies changed over time and determine associated shifts in degrees of policy integration. It would be important to include both policy outputs (i.e. how food security policies have been designed) and outcomes (how they have been implemented) in such an undertaking. Apart from qualitative evidence, the policy integration dimensions could be relatively easily operationalized into quantitative indicators, of course depending on the availability of appropriate data sources. This could, for example, help mapping the range of policies in which food security goals have been adopted (Candel, 2016). For explanatory diagnostics, process-tracing has proved useful for uncovering the causal mechanisms through which a series of effects, here shifts in the policy integration dimensions, come about (Biesbroek et al., 2014; George and Bennett, 2004). Given the limited research on policy-relevant mechanisms (Section 3) such an approach would probably need to be abductive, constantly comparing empirical observations of key actor-interactions with theoretical constructs that can be used for inferring categories of mechanisms (cf. Charmaz, 2006).

Turning to the evaluatory diagnostics, a range of methodological approaches can be used. If the analyst aims to study how scholars and other evaluators have attributed success and failure vis-à-vis food security governance efforts (Section 4), meta-analyses and systematic literature reviews could be performed (e.g., Cooper et al., 2009; Gough et al., 2012). If the analyst aims to assess the degree of success him-/herself, the choice for a methodological approach depends on whether a programmatic or political rationale is applied. When used for a programmatic type of evaluation, methodological approaches are designed to enable collecting and analyzing ‘objective’ data to measure a policy’s effectiveness and/or efficiency in obtaining desired food security outcomes. What makes such analyses challenging is that to determine the causal relationship between (integrated) policy interventions and food security indicators, the analyst will have to control for other, non-policy related, explanations. Two particularly promising ways for doing so are the use of (quasi-)experimental designs and process-tracing methods. Experimental designs allow for determining causal relationships by comparing the effects of an integrated policy intervention on a target group or problem vis-à-vis non- or other types of interventions (for a discussion of various types of experimental designs, see: Knill and Tosun, 2012). In practice, however, scholars are seldom in a position to employ actual macro-policy interventions for such an experimental design (this is different for studies of specific instruments, see, e.g., Stewart et al., 2015). Therefore, they often use quasi-experimental designs, in which the researcher has no control over the substance or the targeting of the intervention. The use of such designs in research on integrated nutrition and child development interventions has shown that exposure to a mix of measures, including nutrition supplements, psychoeducational stimulation, and support to parents, generally has positive effects on children’s cognitive, socioemotional, and/or motoric development (Black et al., 2015). The disadvantage of (quasi-)experimental designs is that they are virtually impossible to use for macro-level integrated policy packages that comprise interventions across a range of sectors and levels. In such a case, the use of process-tracing methods may be more expedient. At the same time, the reliance on qualitative data almost inherently implies that researchers employing these methods will not make the claim that they can objectively measure or accredit the success or failure of an integrated food security strategy. Instead, they will emphasize the plausibility of causal relationships.

A process-tracing methodology can also be used to study how success and failure are constructed in social interactions, i.e. for political evaluation. The goal is then not to uncover causal linkages, but to obtain a better understanding of how stakeholders and recipients experience and appraise an integrated food security strategy and associated range of interventions. Researchers can hereby focus on these interactions and associated institutional dynamics themselves, for example by studying public debates or media coverage, or on sense-making and framing processes of individual actors. Put differently, the researcher’s central concern will be related to the question ‘how does a policy mean?’, i.e. what meanings does it convey?; to what audiences? How is it made sense off? (Yanow, 1996) There is a range of interpretive methods that can be used for such an analysis (e.g., see Fischer, 2003; Wagenaar, 2011; Yanow and Schwartz-Shea, 2006).

6.2 Performing diagnostics in a sub-Saharan african context

The abovementioned lack of descriptive accounts of present-day African governance makes it difficult to raise expectations about what insights would be obtained by performing the three diagnostic steps. In terms of policy design, the quick scan that resulted in the overview in Table 1 showed that some Sub-Saharan African governments (e.g., those of South Africa, Mozambique, and Mali) seem to have developed more comprehensive food security frameworks and on a more sustained basis than others (e.g., those of South Sudan, Equatorial Guinea, or Gabon). At the same time, the case of South Africa’s Integrated Food Security Strategy showed that such designs say little about actual commitments and implementation (Harris et al., 2017). That said, a number of factors may be argued to play a role in whether and how SSA governments decide to adopt integrated food security strategies, how these are being implemented, and whether and how they result in programmatic or political success or failure. It should be noted, however, that these factors are speculative and their relative importance remains unclear.

First, as mentioned in the introduction, the development of many SSA integrated food security strategies has been pushed or supported by international organizations, think tanks and donors, such as FAO, WHO, SUN, IFPRI, and USAID. Many of these actors also monitor the progress that is made in implementing these strategies, possibly putting some pressure on SSA governments to proceed beyond paper realities. It could therefore be hypothesized that SSA countries in which these donors and organizations are relatively more active and on a relatively more sustained basis are more likely to develop and implement (comprehensive) integrated food security strategies.

Second, good governance and effective institutions, including accountable leadership, coordinative capacities, and well-functioning sub-national governmental levels, have been considered key to implementation (Peters and Mawson, 2015; United Nations, 2016). It may therefore be hypothesized that countries that score relatively well in terms of good governance and that have relatively stable and well-functioning governmental institutions are more successful (in programmatic evaluation terms) in implementing food security strategies. Admittedly, it is difficult to measure and systematically compare institutional strength and durability (United Nations, 2016). Analysts therefore often use proxy-indicators derived from large databases such as the Worldwide Governance Indicators (World Bank, 2015), the Corruption Perceptions Index (Transparancy International, 2015), or the Ibrahim Index of African Governance (Mo Ibrahim Foundation, 2015). In addition, it could well be that governments that are experienced in dealing with other crosscutting issues and have consequently developed institutional capacities to deal with these, may be more successful in designing and implementing integrated food security strategies. Many SSA governments have set up national coordinative mechanisms to address HIV/AIDS, malaria, or other infectious diseases.

Third, from a political evaluation perspective, it could be argued that countries with a relatively more stable democratic tradition will be characterized by a more open public debate about the success or failure of public policy, including food security interventions. After all, the possibility for political contestation is a prerequisite for alternative, non-governmental visions to emerge. Somewhat paradoxically, countries that perform relatively well in terms of good governance may therefore also be more susceptible to instances of perceived policy failure.

Fourth, recent comparative research on how states adapt to climate change, an issue with a somewhat similar ‘wicked’ and crosscutting problem nature, found that levels of centralization affect states’ national coordinative capacities, in that centralised states more often install coordinative mechanisms at national level compared to decentralised states (Biesbroek et al., 2016). Building forth on this finding, it could be hypothesized that relatively more centralized SSA countries are more likely to develop (relatively more comprehensive) integrated food security strategies.

Fifth and last, agency obviously matters. Previous research on policy integration has shown that coordinative structures and procedures are important but insufficient if they are not accompanied by considerable and sustained political backing (Peters, 2015). It may be expected that SSA countries of which the leaders and decision-makers have genuinely committed themselves to addressing food insecurity over a sustained period of time are relatively more inclined to develop and implement integrated food security strategies. The mechanismic perspective would enable uncovering the interactions through which such leadership comes to affect policy outcomes.

Again, although based on theoretical insights, these hypotheses and expectations are − by definition − speculative and their applicability and relative importance will need to be assessed by future empirical studies. This paper has aimed to provide some handles for doing so and as such could serve as a point of departure for further research in this field. Most importantly, furthering the state of knowledge about food security interventions in Sub-Saharan Africa would require an increase of scholarly commitment, particularly also from the political sciences, to studying African public governance in general and food governance specifically. I hope this special issue may be an impetus to this pursuit.

Acknowledgments

A previous version of this paper was presented at a workshop of the Netherlands Environmental Assessment Agency (PBL) on ‘Institutional diagnostics for African food security: approaches, methods and implications’, 29–30 September 2016. The author would like to thank Saskia Vossenberg, Sietze Vellema, Martijn Vink and two anonymous reviewers for their helpful comments on previous drafts of this paper.

Notes

1 In this paper, references to ‘food (in)security’ also include concerns about nutrition (security), which is considered a key dimension of food security.

2 This overview was composed in Autumn 2016 through an extensive Google search and by consulting the WHO web page: http://https://extranet.who.int/nutrition/gina/en/policies/summary. NB: I do not claim to be comprehensive; the table should be read as an exploratory overview. For every country, the most recent strategy/-ies that were found are presented.

3 Authors do not explicitly refer to mechanisms, but the (generic) interaction patterns they describe can be considered as such.

4 Authors did not study these mechanisms in the context of policy integration, but to explain impasses in the governance of an innovative climate change adaptation measure.

5 I here synthesize existing (accounts of) evaluations. The analyst could also directly adopt one or both of the evaluation rationales to analyze and appraise an IFSS’s outcomes (see Section 4).

6 A search of ‘integrated food security strategy’ on the newspaper database LexisNexis led to only two relevant articles, both of them criticizing the IFSS’s performance.