Understanding Resilience for Food and Nutrition Security

Several decades ago, short-term shocks were only of peripheral concern to most development experts. Helping people survive natural disasters, like floods and droughts, or manmade ones like civil unrest, was considered the responsibility of humanitarian aid organizations. Conversely, humanitarian agencies have historically focused mainly on relief rather than on the kinds of longer-term development-oriented interventions that might reduce exposure or vulnerability to shocks.

Since then our understanding of the role of short-term shocks has evolved substantially. Even temporary shocks and stressors can have long-term consequences. A poor harvest that reduces a child’s food intake, even temporarily, can have serious effects on her longer-term cognitive and physical development and therefore future earning capacity.

A severe drought that leads a family to sell off its most productive assets, such as its land or livestock, can plunge that family into permanent poverty. It is therefore now widely recognized that a central reason why it is so difficult for poor people to escape poverty is their sheer inability to avoid or cope with shocks and stressors. Yet, at the same time, relief efforts, important though they are, do not typically address the underlying structural vulnerabilities of a population. Recognizing these realities, both the humanitarian and development communities have arrived at a common conclusion: Poor and vulnerable populations need greater resilience, and in order to achieve it, these communities need to work together.

A critical part of building resilience involves boosting food and nutrition security. Poor people have always been vulnerable to “hunger seasons,” droughts, floods, and other natural and man-made disasters (Box 3.1, right and below). In recent years, this perennial vulnerability has been exacerbated by food price and financial crises, and large-scale humanitarian crises such as the recurring droughts and famines in the Sahel and the Horn of Africa.

Several recent crises have even spurred the creation of large-scale programs that explicitly aim to build resilience, including the Global Alliance for Action for Drought Resilience and Growth in the Horn of Africa backed by USAID and the Global Alliance for Resilience in the Sahel (AGIR-Sahel) funded by the European Union (EU). Dozens of other international development projects are being created all over the world to strengthen people’s resilience to shocks and improve their food and nutrition security.

While there is no consensus on the best ingredients for resilience or even its definition, the development and relief communities are clearly moving toward a loosely defined resilience framework that offers the potential for traditionally compartmentalized sectors to design and implement more effective and more integrated interventions. Nevertheless, this emerging resilience framework presents challenges—conceptually, empirically, and practically.

Resilience has roots in the Latin word resilio, meaning “to jump back” (Klein, Nicholls, and Thomalla 2003). Much of the resilience literature broadly defines the term as a return to an original state.

In ecology, resilience has long been concerned with a system’s ability to absorb changes and still persist (Holling 1973). Other resilience studies have focused on the gap between original states and less than ideal conditions. In the 1940s and 1950s, for instance, psychologists studied the negative effects of exclusion, poverty, and traumatic stressors on vulnerable individuals, especially children (Glantz and Johnson 1999). The concept was later adopted in other disciplines, including physics and disaster risk management, with a similar focus on recovery from shocks, or even adverse trends such as rapid population growth.

In the development community, the concept of resilience has been further adapted and elaborated. When applied to complex adaptive systems, resilience is not just about resistance to change and going back to how things were (Folke 2006). It can involve making adjustments to respond to new stresses or even making considerable changes to a system, be it a household, community, or country. Resilience here consists of three capacities that respond to different degrees of change or shocks (Berkes, Colding, and Folke 2003; Walker et al. 2004):

1. Absorptive capacity covers the coping strategies individuals, households, or communities use to moderate or buffer the impacts of shocks on their livelihoods and basic needs.

2. Adaptive capacity is the ability to learn from experience and adjust responses to changing external conditions, yet continue operating.

3. Transformative capacity is the capacity to create a fundamentally new system when ecological, economic, or social structures make the existing system untenable.

According to this broader definition, resilience is the result of not just one, but all three capacities. Each capacity leads to a different outcome: (1) absorptive capacity leads to endurance (or continuity); (2) adaptive capacity leads to incremental adjustments or changes; and (3) transformative capacity leads to transformational, system-changing responses (Figure 3.1).

These three different responses can be linked to different intensities of shock or change in a broadly hierarchical manner. The lower the intensity of the shock, the more likely the household, community, or system will be able to resist it effectively, absorbing its impacts without changing its function, status, or state. For example, a family would be better able to deal with a short-term food price hike—without making drastic changes—than a tsunami that levels its village.

When the shock or stressor exceeds this absorptive capacity, however, individuals and communities will then exercise their adaptive resilience, which involves making incremental changes to keep functioning without major qualitative changes in function or structure. These adjustments can take many forms. Examples include adopting new farming techniques, diversifying one’s livelihood, taking out loans, and connecting to new social networks. These adaptations can be individual or collective, and they can take place at multiple levels, such as among or between households, individuals, or communities.

If, however, those incremental changes associated with adaptive capacity are not enough to prevent a household, community, or system from avoiding dire circumstances, a more substantial transformation must take place. These changes permanently alter the system or structure in question. For example, droughts in the Horn of Africa may push people out of pastoralism and into sedentary agriculture or urban occupations, because they can no longer rebuild their herds (Lybbert et al. 2004; Box 3.2).

Importantly, these changes may not always be positive in the long run, even if they prevent people from falling into acute poverty that puts their access to basic necessities such as food and shelter at risk. In the example described in Box 3.2, those who transition out of pastoralism may fare worse than active pastoralists, since sedentary agriculture is highly risky in arid conditions.

Adopting resilience as an analytical framework could help in the fight against food and nutrition insecurity for several reasons. Resilience helps frame problems coherently and holistically. Linking interrelated short-term shocks and long-term systemic change gives us a more complete view of the factors that lead people to drift into poverty, food and nutrition insecurity, or both. By giving greater weight to the significance of negative shocks than earlier development frameworks did, this concept of resilience highlights how an inability to cope with shocks makes it hard for the poor to escape poverty and explains why others fall into it in the first place (McKay 2009; World Bank 2006).

A resilience framework has practical implications, as well. It may serve as a “mobilizing metaphor” (Béné et al. 2012) to integrate traditionally disparate sectors—particularly the relief and development sectors—and encourage them to work together (USAID 2012). It may also help bolster support for interventions, such as safety-net programs, that bridge relief and development. More integrated multi-sectoral programs and collaborations could adopt a more systemic and holistic approach to fighting both chronic and transient poverty compared with many of today’s piecemeal approaches.

Another practical advantage of using a resilience framework is that it has focused more attention on understanding the welfare and behavioral dynamics of vulnerable populations, including better measurement of transient poverty as well as food and nutrition insecurity.

The analysis and understanding of local dynamics are key to identifying existing and potential self-help competencies and capacities. It is essentially those competencies and capacities that must be built up to increase individuals’, households’, local communities’, and states’ ability to absorb, to adapt, and to transform.

The “resilience lens” thus reaffirms the importance of identifying and strengthening local structures and supporting them in performing their roles effectively and working together. These structures include organizations as diverse as central or decentralized administrations, health centers, disaster risk management committees, and associations of small-scale producers.

While the resilience framework seems to offer many benefits in theory, it faces many challenges in practice. First and foremost, experts in development and humanitarian circles have yet to agree on a common definition of resilience. Too often the definitions adopted tend to emphasize a return to initial states, which hardly seems consistent with promoting transformation and development.

Some critics have also suggested that resilience is a concept that does not translate well from ecological settings to social settings. They argue that the resilience model does not pay enough attention to social dynamics in general, and to issues of agency and power in particular. However, NGOs and other practitioners increasingly challenge this view. They emphasize the resilience-enhancing role played by social processes, such as community cohesion, good leadership, and individual support of collective action (Twigg 2007; Boyd et al. 2008; Schwarz et al. 2011; VFL 2011). A rigorous assessment of the literature shows, however, that the number of these analyses is still low and the evidence thin (Béné et al. 2012).

Others fear that the resilience agenda may be pushed too far, threatening or diluting the impact of more traditional relief activities. If the relief sector’s performance is benchmarked against its contribution to resilience building, many worthwhile but more narrowly focused relief efforts could lose resources. Enthusiasm for resilience building therefore needs to be tempered by an appreciation for the need for core relief activities and the benefits of specialization.

Finally, while resilience usually has positive connotations and is the goal of many programs and projects, the large majority do not consider its possible downsides. Some coping strategies, such as prostitution or begging, may strengthen resilience, but to the detriment of well-being and self-esteem. Other coping activities, such as crime, may increase the resilience of one group to the detriment of another person’s well-being. Moreover, when defined as the rapid return to an initial state, resilience may be counterproductive in the long run. Resilience as “stickiness,” “stubbornness,” or “resistance to change” is clearly not a desirable quality in many circumstances.

These concerns are by no means academic. Populations highly exposed to climate change, such as African pastoralists, are the subject of substantial debate over whether herd recovery or diversification out of pastoralism is the best long-term objective. Similarly, the argument that safety-net programs impede out-migration from drought-prone rural areas is relevant. In such a case, resilience without transformation, in response to a stressor as significant as climate change, could be an undesirable quality in the long run.

As implied, a significant challenge for a resilience framework is to define exactly what value it adds to the current way of doing things. In principle, a resilience framework could add value in two ways. At a strategic level, a resilience framework could encourage governments and development partners to mainstream resilience as a policy and programmatic objective, and to coordinate difference agencies and sectors to achieve that objective. In this strategic sense, it is not obvious that new policy or program instruments are needed to achieve resilience since improved coordination and prioritization could be sufficient in themselves. However, one might also expect a resilience focus to encourage adoption of programs or policies that innovatively bridge the relief and development sectors (as opposed to specializing in one sector or the other).

This raises a question: What types of interventions might build this bridge between relief and development?

An obvious example would be safety-net programs, which meet the criteria for providing social protection, or “relief,” and contributing to development, or “longer-term resilience building.” Social protection typically takes the form of food, cash, or voucher transfers, but the development component is more varied. Transfers that are conditional often incorporate explicit development objectives, such as raising school attendance, expanding vocational training or adult schooling, increasing nutritional knowledge, and, quite commonly, building infrastructure through public works programs.

A very relevant example is the Productive Safety Net Program in Ethiopia (Box 3.3). This program was an innovative solution to two major problems:

(1) the ad hoc, uneven, and unpredictable nature of traditional transfer programs and

(2) the widely held view that excessive focus on relief was inhibiting sustainable rural development.

By combining social protection with public asset building, the Productive Safety Net Program clearly contributes to both relief and longer-term development. In that sense, it is a resilience-oriented program.

Related programs in Ethiopia and elsewhere (such as BRAC’s graduation model in Bangladesh) also focus on helping individuals and households build up business and financial skills as well as confidence and a sense of empowerment. These programs are based on the assumption that providing temporary safety from shocks is a key step toward building up assets that provide a more permanent resilience to shocks.

The Pastoralist Livelihoods Initiative is a quite different example of a relief-and-development intervention from Ethiopia (Box 3.3). While productive safety-net programs are well suited to sedentary crop or crop-livestock systems, pastoralists face unique challenges. Like crops, livestock are highly vulnerable to drought. But unlike annual crops, they are a perennial asset, like land. This makes the death of livestock during droughts potentially very costly. In extreme situations, a household may drop out of pastoralism, simply because it cannot rebuild its herd after a drought.

The Pastoralist Livelihoods Initiative is a tightly focused resilience-building program that switches between relief and development, rather than trying to address both at the same time, as the Productive Safety Net Program does. It offers one clear practical way to address the disconnect between relief and development activities. But while safety-net programs have been widely researched all over the world, more experimentation, learning, and evaluation are required for these kinds of switching programs.

With mounting interest in resilience as a conceptual framework comes increased demand for empirical knowledge of resilience. Governments, nongovernmental organizations, international donors, and others are interested in using the best available indicators and survey instruments to identify differences across space and time and to diagnose sources of vulnerability and design programs to address weaknesses. To diagnose the problems and develop the best responses, it is important to measure resilience by gauging the impacts of both shocks and the mitigating influences on these shocks, such as coping behaviors and outside interventions (Frankenberger and Nelson 2013). In short, good measurement should drive diagnosis and response (Barrett 2010).

A better understanding of resilience will require collecting data on the causes and consequences of a wide range of negative shocks. However, resilience, vulnerability, and coping behaviors are difficult phenomena to measure, because (1) shocks, by definition, are often short-term unpredictable events, implying the need for frequent data (for example, bi-monthly); (2) negative shocks often occur in remote places and populations, such as pastoralists in the Sahel or the Horn of Africa; and (3) resilience to shocks involves complex coping or adaptive behaviors, which are diverse and may involve thresholds and qualitative shifts.

As such, the unpredictable nature of shocks and responses to them makes measuring vulnerability and resilience much more difficult than measuring chronic welfare measures like poverty, child malnutrition, or infant mortality. For chronic measures, occasional snapshots from household surveys usually suffice to paint a general picture of poverty across regions and countries and to determine basic trends. These standard household surveys are not frequent enough, however, to assess the consequences of shocks except by coincidence, and large panel surveys in developing countries are still relatively rare. While many standard economic or health and nutrition surveys might measure important aspects of vulnerability and resilience, they are unlikely to measure all relevant behavioral responses. This suggests that measuring vulnerability and resilience requires a different approach.

What, then, are the key issues that arise when one tries to measure resilience in the context of food and nutrition insecurity? A distinguishing feature of resilience and vulnerability is the potential for complex dynamics. In vulnerable socioeconomic environments, individuals, households, and communities are likely to experience dynamic fluctuations in well-being, including a mix of long-term trends, cyclical and seasonal shocks, and major covariate shocks. Moreover, the transitions from one state, such as chronic poverty, into either better or worse states are likely to be characterized by a range of threshold effects or tipping points, such as when a drought reduces herd sizes below a threshold of recovery (Box 3.2; Lybbert et al. 2004).

Finally, resilience requires a multilevel or systemic measurement approach. This includes measurement at different levels—individual, household, community, (eco)system—and among different socioeconomic and ethnic groups. This also requires an understanding of how these different identities and factors interact. Beyond the household level, systemic factors, such as health conditions, social and political relationships, culture, agroecological factors, and macroeconomic conditions, may affect resilience.

These basic principles have important implications for measurement in practice. Table 3.1 provides a general list of proposed indicators that could be used to measure resilience for food and nutrition security. Perhaps the most important prerequisite for resilience measurement is higher-frequency surveys (Barrett 2010; Headey and Ecker 2013). Though still surprisingly rare, high-frequency measurement is a necessary condition for understanding vulnerability and resilience, because it helps identify (1) “dynamic initial states,” such as seasonality, cyclicality, and exposure to idiosyncratic shocks; (2) differences between pre-shock and post-shock states; (3) the complex dynamics of coping and adaptation mechanisms; and (4) the key thresholds that may arise in the transitions between initial and subsequent states (Barrett and Constas 2012). The more standard program evaluation based on two to three rounds of a survey (typically conducted several years apart) will rarely if ever suffice to make sense of the complexities of highly vulnerable people's lives.

The most pertinent examples of high-frequency resilience surveys are the nutritional surveillance system surveys conducted by Helen Keller International (HKI) in Bangladesh and Indonesia. The World Food Programme (WFP) also uses the nutritional surveillance system approach in some of its high-priority countries, such as South Sudan. These surveys are typically conducted every two months—more often than standard household surveys—in order to pick up the effects of both seasonal shocks and “one-time” natural disasters. Moreover, while the nutritional surveillance system surveys focus heavily on nutrition indicators, they also look at a wide range of household characteristics and coping behaviors (Box 3.4).

Beyond the need to use higher-frequency surveys, resilience measurement faces additional challenges in terms of the breadth of the resilience concept. Resilience is a highly multidimensional concept with numerous causes and manifestations. Moreover, some factors may be considered not only causes or sources of resilience, but also indicators of resilience. For example, a non-exhaustive list of factors that are simultaneously considered as “contributors” to and “results” of resilience includes: technological capacity, appropriate skills and education, gender empowerment, sustainable natural resource management, adequate livelihood assets, good governance, and access to infrastructure (Alinovi et al. 2010; USAID 2012; Tulane and UEH 2012; Vaitla et al. 2012). This clouding of the distinction between cause and effect limits our ability to compare or refute specific hypotheses (Frankenberger and Nelson 2013).

In addition, this diverse and extensive list of factors poses some serious challenges to both measurement and scientific analysis. Some of these factors are inherently difficult to measure, such as governance, natural resource management, and gender empowerment. Many must be measured qualitatively rather than quantitatively. Some indicators must be measured at the individual or household level, but others need to be measured at the community level or even higher. Finally, some factors—as well as the definition of resilience itself—are likely to be context- and shock-specific, thereby limiting comparability across survey sites. Some factors fall under one discipline, such as economics, while others fall under very different disciplines (ecology, political science, sociology). As already emphasized, most—if not all—of these factors ought to be measured in high-frequency surveys. Thus the practical challenges to effectively monitoring and measuring resilience are considerable. Yet collecting such an extensive set of data to measure resilience could help shape more informed responses to a wide range of crises.

The complexity of the concept of resilience and the challenges of measuring and promoting it may paint a somewhat daunting picture for policymakers and development practitioners. Indeed, some vulnerable countries and regions have found themselves mired for decades in poverty and food and nutrition insecurity in the face of shocks. Other highly vulnerable countries, though, have seemingly become more resilient. Much can be learned from the varied experiences of these groups of countries.

Figure 3.2 (below) shows three countries and two sub-regions that score high on the 2013 Global Hunger Index and are exposed to weather shocks, along with their food aid receipts as a proxy for resilience over time. The food aid data reflect the standard narrative of “permanent crisis” in the Sahel and the Horn of Africa, where food aid receipts were roughly as large in 2008–2011 as they were about 20 years ago. In contrast, Malawi and Zambia (two countries where controversial fertilizer subsidy programs have greatly expanded maize production) have seen improvements in recent years, though questions remain about whether these efforts can be sustained. And finally, Bangladesh has achieved a remarkable reduction in food aid dependency. Its 85 percent drop in food aid receipts from the early 1990s to 2008–2011 is consistent with the country’s dramatic economic and social achievements (Economist 2012), including rapid agricultural growth (through new crop varieties and other modern inputs), sharp reductions in fertility rates, dramatic expansion in education (especially for females), a microfinance revolution, and sustained job creation outside of agriculture.

There is more to learn about why some vulnerable regions have made so little progress, while some shock-prone countries seem to have turned themselves around. Success stories like Bangladesh, Malawi, and Zambia, however, show that building individual, community, and national resilience within a generation is a real possibility.