global hunger research paper

An official website of the United States government

Official websites use .gov A .gov website belongs to an official government organization in the United States.

Secure .gov websites use HTTPS A lock ( Lock Locked padlock icon ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites.

Publications
Account settings
Advanced Search
Journal List

The global burden of chronic and hidden hunger revisited: New panel data evidence spanning 1990–2017

Bert lenaerts, matty demont.

Author information
Article notes
Copyright and License information

Corresponding author. Campus Diepenbeek, Agoralaan Building D, 3590, Diepenbeek, Belgium. [email protected]

Received 2020 Jun 15; Revised 2020 Dec 8; Accepted 2020 Dec 9.

This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

Gödecke, Stein and Qaim (2018) (GSQ) recently analysed the 1990–2010 trend and determinants of the global burden of chronic and hidden hunger. We reanalyse and extend GSQ's study and demonstrate that after 2010, significant reductions in the burden of hunger were achieved. In contrast with GSQ, we find that (i) hidden hunger is more prevalent than chronic hunger; (ii) cereal availability and the supply of vegetables and fruits matter; and (iii) gender equality only affects hidden hunger. We further provide evidence on the importance of spatial spillovers in GDP affecting the burden of hunger. Policy makers should therefore prioritize (i) enhancement of micronutrient density of cereals; (ii) diversification of production systems and consumer diets; and (iii) development of nutrition-sensitive food value chains.

Keywords: Hunger, Replication, Spatial spillovers, Cereals, Economic geography, DALYs

We reanalyse and extend the study of Gödecke et al. (2018).

Hidden hunger is more prevalent than chronic hunger.

Supply of cereals, vegetables and fruits significantly alleviates the burden of hunger.

Gender equality only affects hidden hunger and not chronic hunger.

Neighbours' GDP has a major positive effect on the alleviation of the burden of hunger.

1. Introduction

Policymakers face two non-exclusive options to address the burden of hunger: (i) tackle it directly through food security policies; or (ii) counteract its fundamental causes through policies on trade and transport connectivity to overcome geographic disadvantage ( Gallup et al., 1999 ) or policies on agricultural production to overcome unproductive production environments ( Beddington, 2010 ). However, they face two obstacles towards evaluating the global state of hunger and its determinants: (i) lack of country-level data; and (ii) reliance on proxies that only capture selected dimensions of hunger. Gödecke, Stein and Qaim (2018 , 2019 ) (hereafter referred to as GSQ) overcome these problems by employing Disability-Adjusted Life Years (DALYs) from the first generation of the Global Burden of Disease (GBD) dataset quantifying the health burden of hunger in three discrete points in time, that is the years 1990, 2005 and 2010. The authors ﬁnd that (i) between 1990 and 2010, the total burden of chronic and hidden hunger diminished by more than 50 per cent and about 30 per cent, respectively; (ii) in 2010, the global burden of chronic hunger was larger than the global burden of hidden hunger; and (iii) economic growth, urbanisation, democracy, temperate-zone climates, larger food supplies, food diversity, female schooling, and access to improved sanitation and health care have contributed significantly to reducing the burden of chronic and hidden hunger. Remarkably, GSQ found no strong significant effect of cereal availability and the supply of vegetables and fruits on the global burden of chronic and hidden hunger, a result that warrants further investigation. Moreover, while GSQ included determinants of economic performance, demography, institutions and climate, their analysis did not include economic geography, which is an essential determinant of economic development and global food security ( Fujita et al., 1999 ; Gallup et al., 1999 ).

Since the publication by GSQ, the fifth generation of DALY estimates has been released, which are calculated using an updated methodology and span a much broader timeframe (Global Burden of Disease Study 2017) (GBD 2017 GBD 2017 collaborators, 2018 , 2019 ). To generate new evidence on the trends and determinants of the burden of hunger, we revisit the study of GSQ by using new panel data evidence for the period 1990–2017—as an extension and refinement of GSQ's three-point time estimates for 1990, 2005 and 2010—and we enrich the empirical analysis with new variables and alternative model specifications. First, we argue against the use of country-fixed effects panel models given that most factors related to the burden of hunger are fairly time-invariant. Instead, we propose a time-fixed model using two-way clustered standard errors. Secondly, we add determinants of economic geography and finally, we assess how the results are affected by subsampling those countries that are substantially affected by hunger.

2. Global food security and economic development

The global state of food security is largely a reflection of the global state of economic development ( FAO et al., 2019 ). Indeed, global economic development is heterogeneous, as is global food security. Two centuries after the start of the First Industrial Revolution, economic development remains unevenly distributed, with sharp and diverging differences in average income between Western Europe and its offshoots, and the rest of the world ( Bolt et al., 2018 ). While the prevalence of undernourishment in the world has been steadily declining, the total number of undernourished people in Africa, Western Asia and Oceania is larger now than it was a decade ago ( FAO et al., 2019 ). A long-standing question in the literature is how to explain these regional differences in development ( Bhattacharyya, 2016 ). This is important as food security is an outcome of economic development. Traditional economic models emphasise the importance of technological progress (Solow-Swan model; Solow, 1956 ; Swan, 1956 ) and the level of savings and capital accumulation (Harrod-Domar model; Ackley, 1961 ; Baumol, 1970 ). However, these models are problematic because their drivers are largely endogenous (being outcomes of the development process) and therefore need explanation themselves.

Acemoglu (2009) suggests three, non-mutually exclusive, fundamental causes of development: (i) geography, (ii) culture and (iii) institutions and policies. In this paper, we focus mainly on factors related to geography and governance. Geography can be further decomposed into (i) physical geography, (ii) biography, (iii) ecology and (iv) natural resource endowments. Physical geography such as centrality ( Gallup et al., 1999 ), topography ( Nunn and Puga, 2012 ) and landlockedness ( Bloom et al., 1998 ; Demont, 2013 ; Faye et al., 2004 ; Gallup et al., 1999 ; Soullier et al., 2020 ) determine a country's natural level of market potential. This endowed level of market potential can be countered by a country's level of connectivity generated by advances in transport infrastructure and services ( Lenaerts et al., in press ; Pirie, 1993 ; Sheng et al., 2018 ) and trade facilitation ( Brooks and Matthews, 2015 ; Dithmer and Abdulai, 2017 ; Matthews, 2014 ). Both features are an essential part of the concept of market access ( Henderson et al., 2001 ; Lenaerts et al., 2020 ) underlying new economic geography theory ( Krugman, 1991 ). The biogeographic hypothesis states that a large pool of wild plant and animal species suitable for domestication and a continental axis that is oriented east-west (permitting similar ecological zones) have shaped the history of human development in favour of Eurasia ( Diamond, 1997 ). Ecology is understood as the living and production environment, which are characterised by factors such as temperature, rainfall, soil conditions, fragility to degradation and prevalence of pests and diseases (for both plants and animals, including humans) ( Diamond, 2005 ; Gallup et al., 1999 ; Masters and McMillan, 2001 ; Nordhaus, 2006 ; Sachs, 2003 ). These factors both influence the inputs into the production function as well as the production function itself ( Easterly and Levine, 2003 ). Unfavourable ecological conditions for agriculture can be compensated by advancements in agricultural technology ( Beddington, 2010 ; Lenaerts et al., 2019 ). Lastly, natural resources endowments such as oil or metal reserves can, to a certain extent, determine a country's prosperity ( Cavalcanti et al., 2011 ).

Second, cultural factors such as ethnolinguistic fragmentation can affect development, for example through imposing a coordination cost on countries, potentially leading to a negative impact on the quality of government and economic growth ( Easterly and Levine, 1997 ; World Bank, 2008 , p. 104).

Third, institutions and policies can play an important role in fostering development ( Acemoglu and Robinson, 2012 ; Glaeser et al., 2004 ; Rodrik et al., 2004 ). As indicated before, good policies can (partially) counteract the forces behind environmental and cultural determinism. The relationship between institutions, geography and culture is more complicated. Extractive institutions might be historically related to ecology ( Acemoglu et al., 2001 ; Easterly and Levine, 2003 ), while good institutions might interact with both physical geography ( Frankel et al., 1996 ; Frankel and Romer, 1999 ) and the level of political rights for different ethnic groups ( Bloom et al., 1998 , p. 277–278). Traditionally, institutions are captured by metrics assessing (i) the level of democracy, (ii) macroeconomic stability, (iii) good governance (such as political stability and absence of violence) or (iv) the protection of private property ( La Porta et al., 1998 ).

Apart from food security, other outcomes of the development process include demography (such as population (growth) and age distribution ( Bloom et al., 1998 )), agglomeration of firms ( Lenaerts et al., 2020 ) and people (urbanisation) ( Bloem and De Pee, 2017 ; Demont, 2013 ; Headey et al., 2017 ; Satterthwaite et al., 2010 ), and economic performance (such as GDP, employment, productivity, etc.) ( Dollar et al., 2013 ). The outcome of interest in this paper is the state of food security and nutrition in the world, which is affected by other development outcomes and the fundamental causes outlined above ( FAO et al., 2019 ). The state of food security and nutrition can be structured as a hierarchy of causes resulting in undernutrition ( UNICEF, 1991 ): (i) basic determinants discussed above (at the country level); (ii) food security consisting of food availability, food access, food utilisation and stability (at the household level); (iii) hunger, such as hidden hunger, chronic hunger and undernourishment (at the individual level); and (iv) undernutrition (or burden of hunger), such as underweight, stunting and wasting (at the individual level) ( World Food Programme, 2009 ).

3. Methodology

Based on Clemens (2017) , we classify our methodology as a combination of reanalysis (new model specifications) and extension (new data in the form of more years and subsampling)—which are both a type of robustness analysis—of an existing study. GSQ econometrically analysed the drivers of the global burden of chronic and hidden hunger by applying country-level fixed and random effects models on a dataset spread over three points in time. Fixed effects coefficients are typically introduced to control for country-specific characteristics such as culture, history and endowments, which are difficult to measure accurately and change only slowly over time. This reduces bias in the coefficients of the determinants if omitted country-specific characteristics are correlated with the included variables. If the included variables adequately control for country-specific characteristics, the random effects specification is more efficient by correcting for autocorrelation across time, which arises because country-specific characteristics are typically carried over from year to year.

The burden of hunger was measured using Disability-Adjusted Life Years (DALYs) ( IHME., 2013a , IHME., 2013b ). DALYs can capture the outcome of food insecurity, that is, the burden of disease that is caused by hidden and chronic hunger. More speciﬁcally, hunger-related DALYs capture “person-years lost in a population owing to disability and shortened life” ( Stein, 2014 ) as a consequence of adverse health outcomes associated with the burden of hunger. The burden of hunger (measured in DALYs) can be broken down into cause and risk factors. Cause factors are hunger-related diseases or injuries (that is, nutritional deficiencies) that directly cause death or disability; risk factors (per cause-risk attribution) are hunger-related factors that are causally associated with the probability of a disease or injury.

GSQ included seven determinants as fundamental causes for the global burden of hunger, which can be grouped into four categories: economic performance, demography, institutions and climate ( Table 1 ). GSQ also conducted a Hausman test that showed coefficients to differ significantly between the fixed and random effects specifications. This is neither surprising nor very informative since GSQ did not control for country-level geographies like temperature and market potential. As we have argued in Section 2 , geography is a fundamental determinant of economic development and hence food security, which cannot be ignored. Although fixed effects control for country-specific characteristics like geography, they leave no room for interpreting separate effects such as temperature and market potential. Instead, all time-constant characteristics are aggregated at the country-level. A better alternative than merely adding nuisance terms, from which it is difficult to draw any meaningful conclusions, is to attempt to find potential variables that might explain differences in the burden of hunger, including (economic) geography ( Beck, 2001 ).

Overview of basic determinants of the burden of hunger included by GSQ and this study.

Note : ICRG denotes the International Country Risk Guide.

We propose to expand GSQ's set of four categories with a fifth category: economic geography ( Table 1 ; see Table A3 in the supplementary material for more details). To achieve an optimal balance between (i) capturing all categories to minimize omitted variable bias, and (ii) minimizing multicollinearity among the variables, we drop one of the variables in the demographic category, notably the share of youngsters in the population, and keep the variable “per capita GDP” to avoid omitted variable bias, since we expect all drivers of the burden of hunger to be correlated with per capita GDP. This contained variance inflation factors (VIFs) within the range of 1.2–5.7 with a mean VIF of 2.5, which is largely within the acceptable range ( Hair et al., 2011 ; Sheather, 2009 ). To capture institutions, we suggest the International Country Risk Guide (ICRG) risk index instead of the binary democracy indicator employed by GSQ for two reasons: (i) the ICRG index integrates economic, financial and political risk (with political dimensions such as macroeconomic stability, ethnic tensions, socioeconomic conditions, democratic accountability and governance); and (ii) a 100-point scale better captures institutional heterogeneity than a binary indicator. To capture climate, we rely solely on rainfall and temperature as these two are the only variables used to construct the Koeppen-Geiger climate classification. Economic geography is finally captured through trade openness (measured as the share of the value of exports in GDP) and market potential (measured as the inverse-distance weighted sum of GDP for all countries). By employing the fifth generation of the GBD dataset (Global Burden of Disease Study 2017) ( IHME, 2019 ), we update and extend the three-point DALY estimates used by GSQ for 1990, 2005 and 2010 (first generation, Global Burden of Disease Study 2010) to a continuous panel with annual data points spanning 1990–2017 for the descriptive analysis, 1990–2015 for the regression analysis on the basic determinants, 1990–2013 for the regression analysis on food supply and 2000–2015 for the regression analysis on health and gender. Different spellings of country names were matched across data sources. Any remaining missing values were dealt with by omitting the respective year-country record. Further details on the data sources used can be found in the supplementary material.

A second drawback of using (group) fixed effects is that it makes slow-changing variables over time (temporally stable or sluggish variables) to appear either statistically or economically insigniﬁcant ( Beck, 2001 ; Beck and Katz, 2001 ). Unfortunately, many variables of interest related to geography or political economy are almost constant over time. Therefore, “the use of fixed effects will yield odd estimates for coefficients of variables like democracy since the effect of democracy is then “controlled” for [by] the fixed effects” ( Beck and Katz, 2001 , p. 493). To check whether variables are temporally stable, we calculate their level of time invariance. A simple measure of time invariance can be obtained by calculating the R squared of regressing each variable on the set of country-level dummies. High values are an indicator of multicollinearity between the fixed-effects specification and the variables included. Table 2 shows that the level of time invariance for most variables exceeds 95 per cent. This means that after controlling for country-fixed effects only 5 per cent of variance is left to study the impact of the respective variable on the burden of hunger.

A measure of time invariance and probit (first-stage) regression estimates predicting countries with a low burden of hunger.

Notes : GDP, gross domestic product. PPP, purchasing power parity. ***, ** and * denote signiﬁcance at the 1, 5 and 10 per cent level, respectively. Time invariance is calculated as the R squared of a pooled OLS with country-level dummies. For the probit regression, all years between 2000 and 2013 are included.

Another potential problem is the fact that many high-income countries show DALY estimates that are near-zero and therefore are difficult to compare with middle- and low-income countries in terms of the burden of hunger. The burden of hunger in high-income countries is—albeit not zero—marginal compared to that in low-income countries. In addition, by being often located in the Western world, high-income countries differ markedly from middle- and low-income countries. We can formally test this using a binomial regression on a dummy indicating whether a country suffers substantially from hunger ( Blum and Goldfarb, 2006 ). Both hidden and chronic hunger show a heavily right-skewed distribution with a density peak near zero (see Figure A2 in the supplementary material). We use these density peaks as cutoff values to define the set of hunger-afflicted countries: a country is assumed to be free from hunger if its burden of hunger has been below the density peak for both hidden and chronic hunger for every year included in the dataset (see Table A2 in the supplementary material for a list of countries). Using a probit model where the dependent variable is equal to one if the country is not afflicted by hunger, Table 2 indicates that hunger-afflicted countries differ from the rest of the world in terms of per capita GDP, urbanisation, ICRG risk index, average temperature, economic geography, the supply of different types of food, gender equality and public health.

Given these concerns, we suggest (pooled) OLS estimates with clustered (or panel correct) standard errors ( Beck, 2001 ; Beck and Katz, 2001 ). We allow for a small modification by adding yearly fixed effects to control for macroeconomic shocks (such as the 2007–2008 peak in food prices) and calculating two-way clustered standard errors. In addition, we calculate coefficient estimates conditional on countries suffering substantially from hunger (that is, excluding the subsample of hunger-free countries defined above). These conditional effects provide more targeted and relevant insights for policymakers working in middle- and low-income countries.

4.1. Descriptive evidence

Between 1990 and 2017, the total burden of chronic hunger reduced by 72 per cent, while the total burden of hidden hunger declined by 41 per cent ( Fig. 1 ). Expressed per capita, the relative burden of chronic hunger dropped by 80 per cent, while the relative burden of hidden hunger shrunk by 58 per cent. In contrast with GSQ, we find DALY estimates that are 25–50% smaller, particularly because we find chronic hunger to be less prevalent than hidden hunger across the entire timeframe opposite to GSQ. This difference is likely the result of substantial improvements in the estimation of cause and risk factors. For example, while unspeciﬁed anaemia was mapped to iron-deﬁciency anaemia in the Global Burden of Disease Study 2010, this was no longer the case for later generations of the GBD dataset ( GBD 2013 collaborators, 2015 ). In line with GSQ, we find that the burden of chronic hunger has fallen more rapidly than the burden of hidden hunger.

Development of the burden of hunger between 1990 and 2017 by region. (A) DALYs (million) lost due to chronic and hidden hunger globally. (B) DALYs per 1000 capita lost due to chronic and hidden hunger globally. (C) DALYs (million) lost due to chronic hunger regionally. (D) DALYs per 1000 capita lost due to chronic hunger regionally. (E) DALYs (million) lost due to hidden hunger regionally. (F) DALYs per 1000 capita lost due to hidden hunger regionally. Notes : Per capita values represent ratios of global or regional DALYs over population, multiplied by a factor of 1000. SSA, Sub-Saharan Africa. SA, South Asia. EAP, East Asia and Paciﬁc. MNA, Middle East and North Africa. LCR, Latin America and the Caribbean. ECA, Europe and Central Asia. NA, North America. The spike in the EAP curve in C and D is caused by the 1994–1998 North Korean famine. Source: IHME (2019) .

Comparing across regions, we find that Sub-Saharan Africa is the region suffering the most from the burden of chronic hunger, while South Asia features the highest burden of hidden hunger. Expressed per capita, Sub-Saharan Africa clearly suffers more than any other region from both burdens. A similar picture emerges from Fig. 2 , showing a map of 2017 values and changes between 1990 and 2017 in the relative burden of hunger. In 2017, Sub-Saharan Africa and South Asia still suffered substantially more from the relative burden of hunger than the rest of the world (upper quintiles). Other countries in this group include Bolivia, Yemen, Laos and Papua New Guinea.

Map of the relative burden of hunger expressed in quintiles (QU). (A) Chronic hunger per capita in 2017. (B) Hidden hunger per capita in 2017. (C) Relative change in chronic hunger per capita between 1990 and 2017. (D) Relative change in hidden hunger per capita between 1990 and 2017. Source: IHME (2019) .

The largest contributions to the reduction in the global burden of chronic hunger between 1990 and 2017 can be attributed to India (33 per cent), Bangladesh (10 per cent), Ethiopia (6 per cent), China (6 per cent) and Nigeria (5 per cent), while in the case of hidden hunger, the largest contributors to its global alleviation are India (29 per cent), China (14 per cent), Bangladesh (7 per cent), Pakistan (6 per cent) and Indonesia (5 per cent). Although Sub-Saharan Africa and South Asia still suffered the most from the relative burden of hunger in 2017, most of these countries experienced significant reductions in the relative burden of chronic (>75 per cent) and hidden hunger (>50 per cent). However, not every country reported a reduction in the burden of hunger as some high-income countries experienced an increase in the relative burden of chronic hunger. This additionally indicates that care should be taken when assessing the determinants of the burden of hunger for the entire sample of countries worldwide.

Before looking at the regression results, we focus on two factors that affect the burden of hunger and are at the forefront of development policy: per capita GDP and per capita cereal availability (sum of imports and production). The burden of chronic and hidden hunger displays a strongly downward-sloping trend for both factors ( Fig. 3 ). Sub-Saharan Africa features the lowest levels of per capita GDP and per capita cereal availability and correspondingly the highest burden of hunger levels. Countries in East Asia, the Paciﬁc, the Middle East and North Africa exhibit the largest heterogeneity of per capita GDP and cereal availability.

Scatter plot of the burden of hunger (in log, per capita) versus economic performance and cereal availability (sum of imports and production; both in log, per capita) by region. (A) Chronic hunger versus GDP. (B) Hidden hunger versus GDP. (C) Chronic hunger versus the sum of imported and produced cereal. (D) Hidden hunger versus the sum of imported and produced cereal.

Notes : SSA, Sub-Saharan Africa. SA, South Asia. EAP, East Asia and Paciﬁc. MNA, Middle East and North Africa. LCR, Latin America and the Caribbean. ECA, Europe and Central Asia. NA, North America. All years between 1990 and 2017 are included. Source: FAOSTAT (2018) , IHME (2019) , UN DESA (2019) and World Bank (2019) .

4.2. Determinants of the global burden of hunger

Table 3 lists the effect of different basic determinants on the burden of hunger for (i) the full sample of countries (columns 1, 3, 1a and 3a) and the sample of countries suffering substantially from hunger (columns 2, 4, 2a and 4a) and (ii) with (columns 3, 4, 3a and 4a) and without (columns 1, 2, 1a and 2a) controlling for per capita GDP. The results of the regression analysis suggest that per capita GDP (gross domestic product) is still a strong driver of the burden of hunger: a one per cent increase in GDP is associated with a one per cent decrease in the burden of chronic hunger and a 0.6 per cent decrease in the burden of hidden hunger ( Table 3 ). These effects are approximately twice the ones reported by GSQ, indicating that the role of economic growth in alleviating hunger remained important after 2010 ( Dollar et al., 2013 ).

Associations of basic determinants with the burden of chronic and hidden hunger.

Notes : These are results from time-fixed effects panel regressions with annual data from 1990 until 2015 on the subsample of countries listed in Table A2 . Two-way clustered standard errors are shown in parentheses. The dependent variable is the logarithm of DALYs per capita lost due to chronic hunger (models 1–2) and hidden hunger (models 1a–2a). GDP, gross domestic product. PPP, purchasing power parity. ***, ** and * denote signiﬁcance at the 1, 5 and 10 per cent level, respectively.

Further, our results show that urbanisation and good governance have a negative but statistically insignificant effect on the burden of hunger once per capita GDP is controlled for, which is in contrast with the significantly negative effect found by GSQ. A possible explanation for this might be that although urbanisation and governance positively affect food security, they are only indirect effects ( Bezemer and Headey, 2008 ; Ross, 2006 ). The reason why increasing urbanisation and governance appears to lower the burden of hunger significantly in GSQ is that they control for each country's local situation through the use of country-fixed effects. However, on a global scale, governance and urbanisation are only correlated with economic development and therefore, not directly driving eradication of hunger.

In terms of environmental factors, higher levels of rainfall have a statistically significantly negative effect on the burden of hidden hunger while higher temperature levels have a statistically significantly positive effect on both the burden of chronic and hidden hunger. As pointed out by GSQ, this is a troubling finding given that many middle- and low-income countries already face high temperatures and water stress, and climate change is likely to negatively affect many regions in Africa ( Niang et al., 2014 ) and Asia ( Hijioka et al., 2014 ). Although these results are intuitively understood as determinants of the production environment, care should be taken when inferring causality since climate variables like rainfall and temperature are correlated with geography and therefore, might pick up hidden variables. For example, precipitation and temperature peak around the equator but so does ethnolinguistic fragmentation as well ( World Bank, 2008 , p. 104).

To capture economic geography, we included a country's physical centrality to markets (market potential) and openness to trade. Trade openness only appears a statistically significant driver for the burden of hidden hunger for the full sample of countries and when controlling for GDP (column 3a). However, the positive effect is likely a spurious finding ( Brooks and Matthews, 2015 ) as trade openness is no longer statistically significant in the subsample specification (column 4a). Higher levels of market potential are associated with a lower burden of chronic and hidden hunger, although this effect becomes insignificant when we control for GDP and consider the full sample. This implies that a country's standard of living as well as that of its neighbouring countries affect the burden of hunger. Regional clusters in per capita GDP (see Figure A3 in the supplementary material) overlap with the regional trends in the burden of hunger shown in Fig. 2 . Therefore, spatial spillovers in prosperity are one possible explanation for the existence of regional clusters in the burden of hunger. Table 3 indicates that the full sample specification (unconditional effects) differs mainly from the subsample specification (conditional effects) regarding the geographic variables trade openness and market potential. Since we focus on countries facing high burdens of hunger, we are interested in the conditional effects; therefore, we continue with the subsample specification for the remainder of the analysis.

To isolate the effect of the data, we replicated the exact model used by GSQ but with the newer data (see uneven columns in Table A4 in the supplementary material). In comparison with the original results by GSQ, most effects are comparable in size and significance, which points to the improved methodology as the main reason for the differences in regression results in Table 3 with GSQ. Exceptions are the climate variables since the newer data no longer yields a significant effect of average rainfall and only significant effects of climate zones for the burden of chronic hunger (in GSQ, only land area in temperate zones yields significant effects, for both chronic and hidden hunger). The temperate zone effect from GSQ in the estimates of the random effects seems to be captured by average temperature (even columns in Table A4 ), that is, higher temperatures (thus, less temperate zones) are associated with higher burdens of hunger. The effect is reversed in the fixed effects estimates, which can be explained by the low time variance of climate variables such as rainfall and temperature.

4.3. The role of food supply

Table 4 presents evidence on the role of different types of food supply (food availability) on the burden of hunger. Following GSQ, we find that a higher total food supply (kcal/capita/day), in terms of calories from both plant and animal products in general, is associated with a lower burden of chronic and hidden hunger, even after controlling for per capita GDP. Our regression results indicate that all types of animal products, except fish, are negatively associated with the burden of chronic and hidden hunger. In terms of plant products, cereals and pulses matter for the burden of chronic but not hidden hunger. The supply of vegetables and fruits is significantly associated with the burden of chronic and hidden hunger, while there is no statistically significant effect for roots and tubers. These findings for plant products are in line with general views from the field of food science but contrast with the results reported by GSQ (that is, controlling for per capita GDP, they find no effect for cereals and vegetables and fruits, a statistically significantly negative effect for pulses on the burden of hidden hunger and a statistically significantly negative effect for roots and tubers on the burden of chronic and hidden hunger). Our results for cereals are not surprising as cereals are the most important source of calories for the poor but feature low nutritional density in terms of micro-nutrients. Roots and tubers do not have a statistically significant impact on chronic and hidden hunger (when controlled for GDP), perhaps because they mainly provide dietary energy ( Sharma et al., 2016 ). Vegetables and fruits are generally rich in proteins and nutrients explaining their significant impact on chronic and hidden hunger ( Sanchez, 2020 ), although some vegetables such as pulses are rich in proteins rather than micronutrients.

Associations of the food supply with the burden of chronic and hidden hunger.

Notes : These are results from time-fixed effects panel regressions with annual data from 1990 until 2013 on the subsample of countries listed in Table A2 . Two-way clustered standard errors are shown in parentheses. The dependent variable is the logarithm of DALYs per capita lost due to chronic hunger (models 1–6) and hidden hunger (models 1a–6a). GDP, gross domestic product. PPP, purchasing power parity. ***, ** and * denote signiﬁcance at the 1, 5 and 10 per cent level, respectively.

Table 5 provides further evidence by indicating that both higher levels of cereal production and imports, which are a proxy for food availability, help to lower the burden of chronic and hidden hunger. The effect on hidden hunger is smaller (in line with cereals' nutritional composition) but still significant. The latter might be due to the imperfectness of cereal production and imports as proxies for cereal availability, that is, production and import of cereals might be correlated with the production and imports of micronutrient-rich foods. The contribution of cereal imports to reducing the global burden of chronic hunger is almost three times higher than the contribution of domestic cereal production. This highlights the potential role of trade (imports) to supplement the inadequate production of food and hence, strengthen global food security ( Saverimuttu and Rempel, 2004 ). At the same time, however, reliance on food imports makes net food importers vulnerable to market fluctuations such as the 2007–2008 global food crisis ( Moseley et al., 2010 ; Seck et al., 2010 ). After controlling for per capita GDP, both contributions decline and converge. This suggests that it is not trade openness per se but instead purchasing power that affects the burden of hunger, which confirms our previous findings.

Associations of cereals production and imports with the burden of chronic and hidden hunger.

Notes : These are results from time-fixed effects panel regressions with annual data from 1990 until 2013 on the subsample of countries listed in Table A2 . Two-way clustered standard errors are shown in parentheses. The dependent variable is the logarithm of DALYs per capita lost due to chronic hunger. GDP, gross domestic product. PPP, purchasing power parity. ***, ** and * denote significance at the 1, 5 and 10 per cent level, respectively.

4.4. The role of health and gender

Table 6 presents evidence on the role of gender equality and public health (which might affect food access and food utilisation) on the burden of hunger. We find that female school enrolment and the female-to-male life expectancy ratio are both associated with a signiﬁcantly lower burden of hidden hunger. At the same time, access to safe water and improved sanitation and immunisation have a statistically significantly negative effect on the burden of chronic and hidden hunger. In comparison to GSQ, our findings emphasise the importance of public health while only correlating gender equality with the burden of hidden hunger. Given the ongoing Covid-19 pandemic, improving the understanding of the role of public health and gender equality on hunger is a priority towards assessing the impact of the pandemic on middle- and low-income countries. Our results are consistent with the HLPE (2017) food systems framework for diets and nutrition, which explicitly frames women empowerment as a direct driver of consumer behaviour and diets and indirectly through the food environment. The effect of income (GDP per capita) on diets is conditioned by the food environment, which determines the availability, access, affordability, promotion, advertising, information, quality and safety of food ( HLPE, 2017 ). Increased income can reduce hidden hunger where the food environment enables its use on healthy diets, and women's empowerment and gender equality are important mediators of this impact pathway ( Herforth and Ahmed, 2015 ).

Associations of gender equality and public health with the burden of chronic and hidden hunger.

Notes : These are results from time-fixed effects panel regressions with annual data from 2000 until 2015 on the subsample of countries listed in Table A2 . Two-way clustered standard errors are shown in parentheses. The dependent variable is the logarithm of DALYs per capita lost due to chronic hunger (models 1–4) and hidden hunger (models 1a–4a). GDP, gross domestic product. PPP, purchasing power parity. DPT, diphtheria, pertussis, tetanus. ***, ** and * denote signiﬁcance at the 1, 5 and 10 per cent level, respectively.

5. Conclusion

Gödecke et al. (2018 , 2019 ) (GSQ) generated important insights into the trends and determinants of the global burden of chronic and hidden hunger based on a dataset spread over three discrete points in time: 1990, 2005 and 2010. We reanalysed the GSQ study using a more robust specification and extended the data to a panel of countries suffering substantially from hunger, continuously spanning the period 1990–2017. We demonstrate that after 2010, significant reductions in the burden of chronic and hidden hunger were achieved. Since 1990, the relative burdens of chronic and hidden hunger have decreased by 80 and 58 per cent, respectively. Countries typically associated with hunger such as India, Bangladesh and Ethiopia are also among those contributing most to combatting hunger globally. Over the last three decades, global and regional trends show that the burden of hidden hunger has always exceeded the burden of chronic hunger, which contrasts GSQ's earlier findings. In line with GSQ, we find that the burden of chronic hunger is falling more rapidly than the burden of hidden hunger, which suggests that policies should increasingly target hidden hunger.

Following GSQ, we also assessed the impact of country-level determinants on the burden of chronic and hidden hunger. We provide evidence on the importance of spatial spillovers in GDP, which is in line with the existence of supranational clusters in the burden of hunger. Because most countries do not suffer from hunger in isolation, reducing hunger in one country without intervening in neighbouring countries will not suffice to eradicate hunger in the long-run. Development programs with a strong regional focus might be needed. We confirm the effect of per capita GDP and climate in line with GSQ. We find negative but statistically insignificant effects of urbanisation and governance; this might indicate that although both factors can contribute towards lowering the burden of hunger (in line with the literature), on a global scale, they are indirect effects.

Contrasting GSQ, we find that after controlling for per capita GDP, the import and production of cereals and the supply of vegetables (including pulses) and fruits matter in the alleviation of the burden of chronic and hidden hunger; we find no statistically significant impact of root and tubers. These updated conclusions are important as previous findings were not in line with nutritional compositions of the respective food supplies and had far-reaching implications for development policy, especially for cereals. In terms of public health, our results largely follow GSQ, although ours indicate that gender equality only affects hidden hunger and not chronic hunger. Given the ongoing Covid-19 pandemic, improving the understanding of the role of public health and gender equality on hunger is a priority towards assessing the impact of the pandemic on middle- and low-income countries.

Most determinants have larger eﬀects on the burden of chronic hunger than on the burden of hidden hunger, which corroborates the steeper decline in the burden of chronic hunger. As a result, hidden hunger has surpassed chronic hunger over the last three decades. This means that policy makers should prioritize (i) enhancement of micronutrient density of cereals through biofortification and industrial fortification; (ii) diversification of production systems and consumer diets; and (iii) development of nutrition-sensitive food value chains, in which the food environment, women's empowerment and gender equality play crucial mediating roles.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

Financing from the Bill & Melinda Gates Foundation, Seattle, WA, USA [Grant no. OPP1194925] and the CGIAR Research Program on Rice is gratefully acknowledged. The authors would like to thank Theda Gödecke and Alexander J. Stein for their helpful insights. Finally, we would like to thank the editor and two anonymous reviewers for their constructive comments that greatly improved this article. The authors attest originality of the results and assume full responsibility for any remaining shortcomings.

Supplementary material related to this article can be found at https://doi.org/10.1016/j.gfs.2020.100480 .

Appendix A. Supplementary material

The following is the supplementary material related to this article:

Acemoglu D. Princeton University Press; Princeton, New Jersey: 2009. Introduction to Modern Economic Growth. [ Google Scholar ]
Acemoglu D., Johnson S., Robinson J.A. The colonial origins of comparative development: an empirical investigation. Am. Econ. Rev. 2001;91(5):1369–1401. [ Google Scholar ]
Acemoglu D., Robinson J.A. Crown Publishers; New York: 2012. Why Nations Fail: The Origins of Power, Prosperity, and Poverty. [ Google Scholar ]
Ackley G. Macroeconomic Theory. Macmillan; New York: 1961. Economic Growth: The Problem of Capital Accumulation; pp. 505–535. [ Google Scholar ]
Baumol W.J. Economic Dynamics: an Introduction. Macmillan; London: 1970. Mr. Harrod's model; pp. 37–55. [ Google Scholar ]
Beck N. Time-series–cross-section data: what have we learned in the past few years? Annu. Rev. Polit. Sci. 2001;4(1):271–293. [ Google Scholar ]
Beck N., Katz J.N. Throwing out the baby with the bath water: a comment on Green, Kim, and Yoon. Int. Organ. 2001;55(2):487–495. [ Google Scholar ]
Beddington J. Food security: contributions from science to a new and greener revolution. Philos. Trans. R. Soc. Lond. B Biol. Sci. 2010;365(1537):61–71. doi: 10.1098/rstb.2009.0201. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ]
Bezemer D., Headey D. Vol. 36. World development; 2008. Agriculture, development, and urban bias; pp. 1342–1364. 8. [ Google Scholar ]
Bhattacharyya S. The Oxford Handbook of the Social Science of Poverty. 2016. The historical origins of poverty in developing countries; p. 270. [ Google Scholar ]
Bloem S., De Pee S. How urbanization patterns can Guide strategies for achieving adequate nutrition. In: De Pee S., Taren D., Bloem M.W., editors. Nutrition and Health in a Developing World. Springer; 2017. [ Google Scholar ]
Bloom D.E., Sachs J.D., Collier P., Udry C. Geography, demography, and economic growth in Africa. Brookings Pap. Econ. Activ. 1998;1998(2):207–295. [ PubMed ] [ Google Scholar ]
Blum B.S., Goldfarb A. Does the internet defy the law of gravity? J. Int. Econ. 2006;70(2):384–405. [ Google Scholar ]
Bolt J., Inklaar R., de Jong H., van Zanden J.L. 2018. Maddison Project Database, Version 2018, “Rebasing ‘Maddison’: New Income Comparisons and the Shape of Long-Run Economic Development”, Maddison Project Working Paper 10. [ Google Scholar ]
Brooks J., Matthews A. 2015. Trade Dimensions of Food Security, Advance Access Published 2015. [ Google Scholar ]
Cavalcanti T.V. de V., Mohaddes K., Raissi M. Growth, development and natural resources: new evidence using a heterogeneous panel analysis. Q. Rev. Econ. Finance. 2011;51(4):305–318. [ Google Scholar ]
Clemens M.A. The meaning of failed replications: a review and proposal. J. Econ. Surv. 2017;31(1):326–342. [ Google Scholar ]
Demont M. Reversing urban bias in African rice markets: a review of 19 national rice development strategies. Glob. Food Secur. 2013;2(3):172–181. [ Google Scholar ]
Diamond J. W. W. Norton & Company; New York: 1997. Guns, Germs, and Steel: The Fates of Human Societies. [ Google Scholar ]
Diamond J. Penguin; New York: 2005. Collapse: How Societies Choose to Fail or Succeed. [ Google Scholar ]
Dithmer J., Abdulai A. Does trade openness contribute to food security? A dynamic panel analysis. Food Pol. 2017;69:218–230. [ Google Scholar ]
Dollar D., Kleineberg T., Kraay A. The World Bank; 2013. Growth Still Is Good for the Poor. [ Google Scholar ]
Easterly W., Levine R. Africa's growth tragedy: policies and ethnic divisions. Q. J. Econ. 1997;112(4):1203–1250. [ Google Scholar ]
Easterly W., Levine R. Tropics, germs, and crops: how endowments influence economic development. J. Monetary Econ. 2003;50(1):3–39. [ Google Scholar ]
FAOSTAT. 2018. http://faostat.fao.org [ Google Scholar ]
FAO. IFAD. UNICEF. WFP. WHO . FAO; Rome: 2019. The State of Food Security and Nutrition in the World 2019. Safeguarding against Economic Slowdowns and Downturns. [ Google Scholar ]
Faye M.L., McArthur J.W., Sachs J.D., Snow T. The challenges facing landlocked developing countries. J. Hum. Dev. 2004;5(1):31–68. [ Google Scholar ]
Frankel J.A., Romer D.H. Does trade cause growth? Am. Econ. Rev. 1999;89(3):379–399. [ Google Scholar ]
Frankel J.A., Romer D., Cyrus T. National Bureau of Economic Research; 1996. Trade and Growth in East Asian Countries: Cause and Effect? [ Google Scholar ]
Fujita M., Krugman P.R., Venables A.J. Wiley Online Library; 1999. The Spatial Economy: Cities, Regions and International Trade. [ Google Scholar ]
Gallup J.L., Sachs J.D., Mellinger A.D. Geography and economic development. Int. Reg. Sci. Rev. 1999;22(2):179–232. [ Google Scholar ]
GBD 2013 collaborators Global, regional, and national age-sex specific all-cause and cause-specific mortality for 240 causes of death, 1990-2013: a systematic analysis for the Global Burden of Disease Study 2013. Lancet. 2015;385(9963):117–171. doi: 10.1016/S0140-6736(14)61682-2. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ]
GBD 2017 collaborators Global, regional, and national disability-adjusted life-years (DALYs) for 359 diseases and injuries and healthy life expectancy (HALE) for 195 countries and territories, 1990–2017: a systematic analysis for the Global Burden of Disease Study 2017. Lancet. 2018;392(10159):1859–1922. doi: 10.1016/S0140-6736(18)32335-3. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ]
GBD 2017 collaborators Health effects of dietary risks in 195 countries, 1990–2017: a systematic analysis for the Global Burden of Disease Study 2017. Lancet. 2019;393(10184):1958–1972. doi: 10.1016/S0140-6736(19)30041-8. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ]
Glaeser E.L., La Porta R., Lopez-de-Silanes F., Shleifer A. Do institutions cause growth? J. Econ. Growth. 2004;9(3):271–303. [ Google Scholar ]
Godecke T., Stein A.J., Qaim M. The global burden of chronic and hidden hunger: trends and determinants. Glob. Food Secur. 2018;17:21–29. [ Google Scholar ]
Godecke T., Stein A.J., Qaim M. Corrigendum to ‘The global burden of chronic and hidden hunger: trends and determinants’. Glob. Food Secur. 2019;22:46. [ Google Scholar ]
Hair J.F., Ringle C.M., Sarstedt M. PLS-SEM: indeed a silver bullet. J. Market. Theor. Pract. 2011;19(2):139–152. [ Google Scholar ]
Headey D., Stifel D., You L., Guo Z. 2017. Remoteness, Urbanization and Child Nutrition in Sub‐saharan Africa, Agricultural Economics, Advance Access Published 2017. [ Google Scholar ]
Henderson J.V., Shalizi Z., Venables A.J. Geography and development. J. Econ. Geogr. 2001;1(1):81–105. [ Google Scholar ]
Herforth A., Ahmed S. The food environment, its effects on dietary consumption, and potential for measurement within agriculture-nutrition interventions. Food Secur. 2015;7(3):505–520. [ Google Scholar ]
Hijioka Y., Lin E., Pereira J.J., Corlett R.T., Cui X., Insarov G.E., Lasco R.D., Lindgren E., Surjan A. Asia. In: Barros V.R., Field C.B., Dokken D.J., Mastrandrea M.D., Mach K.J., Bilir T.E., Chatterjee M., Ebi K.L., Estrada Y.O., Genova R.C., Girma B., Kissel E.S., Levy A.N., MacCracken S., editors. Climate Change 2014: Impacts, Adaptation, and Vulnerability. Part B: Regional Aspects. Contribution of Working Group II to the Fifth Assessment Report of the Intergovernmental Panel of Climate Change. Cambridge University Press; Cambridge, United Kingdom and New York: 2014. pp. 1327–1370. [ Google Scholar ]
HLPE . 2017. Nutrition and Food Systems. A Report by the High Level Panel of Experts on Food Security and Nutrition of the Committee on World Food Security, Rome. [ Google Scholar ]
IHME. IHME; Seattle, United States: 2013. Global Burden of Disease Study 2010 (GBD 2010): Results by cause 1990–2010. http://ghdx.healthdata.org/record/global-burden-disease-study-2010-gbd-2010-results-cause-1990-2010-country-level Available from. [ Google Scholar ]
IHME. IHME; Seattle, United States: 2013. Global Burden of Disease Study 2010 (GBD 2010): Results by risk factor 1990–2010. http://ghdx.healthdata.org/record/global-burden-disease-study-2010-gbd-2010-results-risk-factor-1990-2010 Available from. [ Google Scholar ]
IHME . IHME; Seattle, United States: 2019. Global burden of disease study 2017 (GBD 2017) results. Available from. [ DOI ] [ Google Scholar ]
Krugman P. Increasing returns and economic geography. J. Polit. Econ. 1991;99(3):483–499. [ Google Scholar ]
La Porta R., Lopez-de-Silanes F., Shleifer A., Vishny R.W. Law and finance. J. Polit. Econ. 1998;106(6):1113–1155. [ Google Scholar ]
Lenaerts B., Allroggen F., Malina R. The economic impact of aviation: a review on the role of market access. J. Air Transport. Manag. 2020 doi: 10.1016/j.jairtraman.2020.102000. (in press). Available from. [ DOI ] [ Google Scholar ]
Lenaerts, B., Allroggen, F., Malina, R., in press. Measuring the quality of air transport networks – a topology of connectivity and accessibility metrics. Handbook Air Transp. Reg. Develop. Taylor and Francis, London.
Lenaerts B., Collard B.C., Demont M. Improving global food security through accelerated plant breeding. Plant Sci. 2019;287 doi: 10.1016/j.plantsci.2019.110207. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ]
Masters W.A., McMillan M.S. Climate and scale in economic growth. J. Econ. Growth. 2001;6(3):167–186. [ Google Scholar ]
Matthews A. Trade rules, food security and the multilateral trade negotiations. Eur. Rev. Agric. Econ. 2014;41(3):511–535. [ Google Scholar ]
Moseley W.G., Carney J., Becker L. Neoliberal policy, rural livelihoods, and urban food security in West Africa: a comparative study of the Gambia, Côte d'Ivoire, and Mali. Proc. Natl. Acad. Sci. Unit. States Am. 2010;107(13):5774–5779. doi: 10.1073/pnas.0905717107. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ]
Niang I., Ruppel O.C., Abdrabo M.A., Essel A., Lennard C., Padgham J., Urquhart P. Africa, pp. 1199–1265. In: Barros V.R., Field C.B., Dokken D.J., Mastrandrea M.D., Mach K.J., Bilir T.E., Chatterjee M., Ebi K.L., Estrada Y.O., Genova R.C., Girma B., Kissel E.S., Levy A.N., MacCracken S., editors. Climate Change 2014: Impacts, Adaptation, and Vulnerability. Part B: Regional Aspects. Contribution of Working Group II to the Fifth Assessment Report of the Intergovernmental Panel of Climate Change. Cambridge University Press; Cambridge, United Kingdom and New York: 2014. [ Google Scholar ]
Nordhaus W.D. Geography and macroeconomics: new data and new findings. Proc. Natl. Acad. Sci. Unit. States Am. 2006;103(10):3510–3517. doi: 10.1073/pnas.0509842103. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ]
Nunn N., Puga D. Ruggedness: the blessing of bad geography in Africa. Rev. Econ. Stat. 2012;94(1):20–36. [ Google Scholar ]
Pirie G.H. Transport, food insecurity and food aid in sub-Saharan Africa. J. Transport Geogr. 1993;1(1):12–19. [ Google Scholar ]
Rodrik D., Subramanian A., Trebbi F. Institutions rule: the primacy of institutions over geography and integration in economic development. J. Econ. Growth. 2004;9(2):131–165. [ Google Scholar ]
Ross M. Is democracy good for the poor? Am. J. Polit. Sci. 2006;50(4):860–874. [ Google Scholar ]
Sachs J.D. National Bureau of Economic Research; 2003. Institutions Don't Rule: Direct Effects of Geography on Per Capita Income. [ Google Scholar ]
Sanchez P.A. Time to increase production of nutrient-rich foods. Food Pol. 2020;91 [ Google Scholar ]
Satterthwaite D., McGranahan G., Tacoli C. Urbanization and its implications for food and farming. Philos. Trans. R. Soc. Lond. B Biol. Sci. 2010;365(1554):2809–2820. doi: 10.1098/rstb.2010.0136. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ]
Saverimuttu V., Rempel H. The determinants of cereal grain imports: sub‐Saharan Africa, 1970–1997. Afr. Dev. Rev. 2004;16(3):525–548. [ Google Scholar ]
Seck P.A., Tollens E., Wopereis M.C., Diagne A., Bamba I. Rising trends and variability of rice prices: threats and opportunities for sub-Saharan Africa. Food Pol. 2010;35(5):403–411. [ Google Scholar ]
Sharma H.K., Njintang N.Y., Singhal R.S., Kaushal P. John Wiley & Sons; 2016. Tropical Roots and Tubers: Production, Processing and Technology. [ Google Scholar ]
Sheather S. Springer Science & Business Media; New York, USA: 2009. A Modern Approach to Regression with R. [ Google Scholar ]
Sheng Y., Jackson T., Lawson K. Evaluating the benefits from transport infrastructure in agriculture: a hedonic analysis of farmland prices. Aust. J. Agric. Resour. Econ. 2018;62(2):237–255. [ Google Scholar ]
Solow R.M. A contribution to the theory of economic growth. Q. J. Econ. 1956;70(1):65–94. [ Google Scholar ]
Soullier G., Demont M., Arouna A., Lançon F., del Villar P.M. The state of rice value chain upgrading in West Africa. Glob. Food Secur. 2020;25:100365. doi: 10.1016/j.gfs.2020.100365. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ]
Stein A.J. Rethinking the measurement of undernutrition in a broader health context: should we look at possible causes or actual effects? Glob. Food Secur. 2014;3(3):193–199. [ Google Scholar ]
Swan T.W. Economic growth and capital accumulation. Econ. Rec. 1956;32(2):334–361. [ Google Scholar ]
UN DESA 2019. https://population.un.org/wpp/Download/Standard/Population
UNICEF Strategy for improved nutrition of children and women in developing countries. Indian J. Pediatr. 1991;58:13–24. doi: 10.1007/BF02810402. [ DOI ] [ PubMed ] [ Google Scholar ]
World Bank . The World Bank; Washington, DC: 2008. World Development Report 2009: Reshaping Economic Geography. [ Google Scholar ]
World Bank 2019. https://data.worldbank.org
World Food Programme . 2009. World Hunger Series, Hunger and Markets. [ Google Scholar ]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

View on publisher site
Collections

Add to Collections

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

View all journals
Explore content
About the journal
Publish with us
Sign up for alerts
30 August 2021

Food systems: seven priorities to end hunger and protect the planet

Joachim von Braun 0 ,
Kaosar Afsana 1 ,
Louise O. Fresco 2 &
Mohamed Hassan 3

Joachim von Braun is professor for economic and technological change at Bonn University, Bonn, Germany; and chair of the Scientific Group for the UN Food Systems Summit 2021.

You can also search for this author in PubMed Google Scholar

Kaosar Afsana is professor of public health in the James P. Grant School of Public Health, BRAC University, Dhaka, Bangladesh; and vice-chair of the Scientific Group for the UN Food Systems Summit 2021.

Louise O. Fresco is president of the executive board, Wageningen University & Research, Wageningen, the Netherlands; and vice-chair of the Scientific Group for the UN Food Systems Summit 2021.

Mohamed Hassan is president of The World Academy of Sciences (TWAS), Trieste, Italy; and vice-chair of the Scientific Group for the UN Food Systems Summit 2021.

School children in Madagascar eat lunch provided as part of a nutrition initiative run by the World Food Programme. Credit: Rijasolo/AFP/Getty

You have full access to this article via your institution.

The world’s food system is in disarray. One in ten people is undernourished. One in four is overweight. More than one-third of the world’s population cannot afford a healthy diet. Food supplies are disrupted by heatwaves, floods, droughts and wars. The number of people going hungry in 2020 was 15% higher than in 2019, owing to the COVID-19 pandemic and armed conflicts 1 .

Our planetary habitat suffers, too. The food sector emits about 30% of the world’s greenhouse gases. Expanding cropland, pastures and tree plantations drive two-thirds of the loss in forests (5.5 million hectares per year), mostly in the tropics 2 . Poor farming practices degrade soils, pollute and deplete water supplies and lower biodiversity.

As these interlinkages become clear, approaches to food are shifting — away from production, consumption and value chains towards safety, networks and complexity. Recent crises around global warming and COVID-19 have compounded concerns. Policymakers have taken note.

Counting the hidden $12-trillion cost of a broken food system

In September, the United Nations secretary-general, António Guterres, will convene a Food Systems Summit. This is only the sixth UN summit on food since 1943, and the first with heads of states in the UN General Assembly. A group of leading scientists, known as the Scientific Group, has been tasked with ensuring that the science underpinning the 2021 summit is robust, broad and independent. We, the authors, are this group’s chair and vice-chairs. Although such approaches are familiar in areas such as climate change and biodiversity, this marks the first time that scientists have been explicitly brought in to multilateral discussions around food (see Nature 595 , 332; 2021 ).

The global food system needs a revamp — in policies and institutions as well as on social, business and technology fronts 3 . Science is one lens for making sure that changes are integrated and collectively deliver better outcomes. But the task is challenging. Food spans many disciplines — not least agriculture, health, climate science, artificial intelligence and digital science, political science and economics. The indirect, adverse effects of policies on climate change, biodiversity loss and health need to be factored in to the true costs of food; these are estimated to be about twice the current market value of food consumption globally 4 . A range of voices is crucial. The Scientific Group is engaging with hundreds of experts across civil society, including Indigenous peoples, producer and youth organizations and the private sectors.

Here we highlight the key roles that scientists should take to accelerate the transformation to healthier, more sustainable, equitable and resilient food systems. These seven priorities reflect the Scientific Group’s evidence base, comprising more than 50 reports and briefs (see go.nature.com/3dtoazu ).

A Bolivian woman in traditional dress carrying a bushel of quinoa in a field

A farmer in Bolivia harvests quinoa for its edible seeds. Credit: Giraudou Laurent/Sygma/Getty

Seven priorities

Science-driven advances are needed to address the following challenges.

End hunger and improve diets. Scientists need to identify optimal conditions and opportunities for investments to make healthy and nutritious foods more available, affordable and accessible. Measures that jointly improve more than one of these are most effective. For example, increased irrigation on small farms in Tanzania and Ethiopia has enhanced productivity, dietary diversity and farmers’ incomes 5 .

Three big game-changers are: enhancing research and development in agriculture and food to increase productivity in a sustainable way; slashing food waste and losses; and adding income and nutrition components to social-protection programmes. Research priorities to cut waste include scaling up solar energy and battery storage technologies to make food processing and preservation more affordable. New forms of packaging using recycled materials, coatings of nanomaterials and even edible films would keep foods fresh for longer. School feeding programmes, together with incentives to keep children in education (such as take-home rations for parents) have seen success in Mali, for example, where they increased school enrolment by 10 percentage points 6 . Under COVID-19 lockdowns, these types of programme became even more relevant.

Europe’s Green Deal offshores environmental damage to other nations

Researchers also need to study behavioural barriers to healthy eating, such as snacking under stress. They should develop policy guidelines for educational food labels, and model the impacts of putting taxes and regulations on unhealthy foods (such as sugar and trans-fats). The health properties of fortified foods and cultivated meats must also be established.

De-risk food systems. The more global, dynamic and complex food systems become, the more open they are to new risks. Scientists need to improve how they understand, monitor, analyse and communicate such vulnerabilities. For example, droughts, the expansion of biofuels and financial speculation after the sudden imposition of trade barriers led to food price hikes in 2008 7 . The COVID-19 pandemic and armed conflicts have shaken food value chains across Africa this year, driving up food prices. Successful initiatives exist, combining on-the-ground observations of food systems and nutrition with forecasting. These include FEWS NET ( https://fews.net ) and the joint analyses from the UN Food and Agriculture Organization and the World Food Programme on early warnings of food insecurity 8 .

Policies and economic solutions are needed. For example, new insurance products aided by remote sensing and weather forecasts would provide cover for lost crops and livestock. Solar-powered irrigation systems would reduce risk from drought. Smartphone apps would provide farmers with information on local crop pests, weather risks and market opportunities; these are already used in Kenya, Senegal, India and Bangladesh 9 . Payment schemes are needed to encourage farmers to manage and capture carbon in soils and trees, and to trade it.

A female seaweed farmer stands in the shallows holding a mass of seaweed in Indonesia

A seaweed farmer in Bali, Indonesia. Credit: Anton Raharjo/NurPhoto/Getty

Protect equality and rights. Poverty and inequalities associated with gender, ethnicity and age restrict many people’s access to healthy foods. Socio-economic researchers need to suggest inclusive ways to transform the more than 400 million smallholder farms worldwide. They must identify pathways out of inequitable and unfair arrangements over land, credit and labour, and empower the rights of women and youth. For example, if female-headed households in southern Ethiopia had the same resources as male-headed ones, their productivity in maize (corn) would increase by more than 40%, to match that of the latter 10 .

Protecting the land rights of smallholders, women and Indigenous peoples is paramount. Technology can ensure transparency and efficiency. For example, using blockchain ledgers of ownership rights to allocate land could be an opportunity in Ghana 11 . At the trans-national scale, the Land Matrix Initiative collects and shares data on big land acquisitions and investments in low- and middle-income countries; it covers deals in almost 100 nations worldwide. Similar solutions are needed to protect the land rights of Indigenous peoples 12 . Also required are efforts to build local research capacity, educational programmes around food and farming, and training and financing opportunities in rural areas.

Boost bioscience. Researchers need to find ways to restore soil health and improve the efficiency of cropping, crop breeding and recarbonizing the soil and biosphere. Linkages among all Earth systems must be considered together — known as a One Health approach (see go.nature.com/3jy7ekh ).

Five priorities for a sustainable ocean economy

Alternative sources of healthy protein need to be advanced, such as plant-based and insect-derived proteins, including for animal feed. Plant-breeding techniques that capture nitrogen from the air, to reduce the need for fertilizers and increase nutrients, should be investigated. Genetic engineering and biotechnology should be applied to increase the productivity, quality and resistance of crops to pests and drought. Recent examples include banana varieties that are resistant to Fusarium wilt fungal diseases, and pest-resistant Bt aubergines. To widen access to bioscience technologies, intellectual-property rights, skills and data sharing should be addressed.

Protect resources. Tools are needed to help people to manage soils, land and water sustainably. For example, hand-held digital devices and remote sensing can track concentrations of soil carbon and other nutrients. Artificial-intelligence systems and drones allow farmers to spot areas that need irrigation, fertilization and protection from pests. Soil microbes can be harnessed to improve soil structure, carbon storage and yields. Researchers need to adapt and scale up such technologies.

Biodiversity and genetic bases need to be protected. Seed varieties must be preserved, and their phenotypes and genotypes explored in the contexts of climate change and nutrition. Traditional food and forest systems, including those of Indigenous peoples, need to be better understood and supported in national agricultural research systems. Cooperation for mutual benefit should be explored, as has been done for climate adaptation in US Indigenous areas 13 .

Averting hunger in sub-Saharan Africa requires data and synthesis

Sustain aquatic foods. Most of the focus on food so far has been on soil-based agriculture. Fish, shellfish and aquatic plants such as seaweed have much to offer nutritionally and environmentally. Aquatic foods need to be better integrated into the understanding of food systems 14 . Researchers should look for ways to increase nutritional diversity in aquatic foods and sequester carbon in marine and freshwater environments.

Ecological-science perspectives and global cooperation and institutions are needed to make harvesting of oceans, coastal waters and freshwater resources sustainable and protect biodiversity. The sustainability of fish-feeding systems needs attention; for example, by exploring the use of insect rearing, oil-rich modified legumes and micro-algae as fish feed.

Harness digital technology. Robots, sensors and artificial intelligence are increasingly used on farms: to harvest crops and milk cows, for example. Sensors can monitor the origin and quality of ingredients and products along the food-processing chain to reduce losses and guarantee food safety. But most farmers and producers still don’t have access. To spread the benefits, devices need to become cheaper and easier to purchase and use. Rental services similar to Uber for farm machinery should be developed, as has been done with tractors in India. Rural electricity supplies will have to be expanded, along with IT training and education.

First steps

The 2021 Food Systems Summit is a great opportunity to end hunger by 2030 and set in train a sustainable food system. Previous summits have delivered change: creating the FAO (after 1943); strengthening the global food-research partnership CGIAR and founding the International Food Policy Research Institute (after 1974); accelerating the human right to food (2002); and establishing monitoring systems to warn of food-price crises (2009).

Without food, there can be no exit from the pandemic

The breadth of the 2021 agenda could be a hindrance, however, to achieving its goals. To avoid failure, delegates should focus. They should prioritize establishing a guiding framework — for transforming diverse national and local food systems, as well as global networks, with the challenges of trade, finance, climate, innovation and governance.

Debates will be fierce. Food is a contentious topic. Disagreements abound over goals, pathways and speed of change, and the roles of science and technology, the private sector and the UN. For example, some see agroecology as the only acceptable way of farming; some view biotechnology and gene editing as dangers; others see opportunities. The Scientific Group has aimed to offer a scientific basis to this diversity of perspectives.

Actions and targets

Once plans are agreed, the UN Food Systems Summit will need to move to implementation. Here are our suggestions.

Nature-based solutions can help cool the planet — if we act now

First, boost finance. On the research front, we propose that governments allocate to food research at least 1% of the proportion of their nations’ gross domestic product that relates to food systems. Many countries spend only half of that. Least-developed countries should be given aid to reach a similar level. To end hunger for the poorest, we propose the establishment of a special fund. This would be supported by development-aid donors and bonds backed by the International Monetary Fund and World Bank. Research and modelling would be required on implementation and impacts.

Second, increase scientific capacity. The special fund could be used to strengthen research capacity in low- and middle-income countries, and expand research collaborations between the public and private sectors, and among farmers, start-up firms in food value chains and science communities. Sharing research infrastructure and data between the global south and global north would be a good start.

Third, strengthen science–policy interfaces. In stark contrast to many other fields, agriculture, food security and nutrition do not have an international agreement or convention to consolidate actions. We call on the UN Food Systems Summit and UN member states to explore an intergovernmental treaty or framework convention on food systems, analogous to the conventions on climate, biodiversity and desertification agreed on in Rio de Janeiro in 1992. The framework will need to include a strong independent scientific body that provides policy advice in the follow-up to the summit. We recommend that all science organizations and academies with food-relevant research be included in a preparatory process.

Bringing the tools of science to the table will help to transform the global food system to end hunger and achieve the UN Sustainable Development Goals by 2030.

Nature 597 , 28-30 (2021)

doi: https://doi.org/10.1038/d41586-021-02331-x

Food & Agriculture Organization, International Fund for Agricultural Development, UNICEF, World Food Programme and World Health Organization. The State of Food Security and Nutrition in the World 2021 (FAO, 2021).

Google Scholar

Pendrill, F., Persson, U. M., Godar, J. & Kastner, T. Environ. Res. Lett. 14 , 055003 (2019).

Article Google Scholar

Organisation for Economic Co-operation and Development. Making Better Policies for Food Systems (OECD, 2021).

Hendriks, S. et al. The True Cost and True Price of Food (United Nations, 2021).

Passarelli, S., Mekonnen, D., Bryan, E. & Ringler, C. Food Sec . 10 , 981–997 (2018).

Aurino, E., Tranchant, J.-P, Diallo, A. S. & Gelli, A. J . Dev. Stud. 55 (Suppl. 1), 7–28 (2019).

Kalkuhl, M., von Braun, J. & Torero, M. (eds) Food Price Volatility and Its Implications for Food Security and Policy (Springer, 2016).

Food & Agriculture Organization & World Food Programme. FAO-WFP Early Warning Analysis of Acute Food Insecurity Hotspots, July 2020 (FAO & WFP, 2020).

Baumüller, H. In Handbook on ICT in Developing Countries: 5G Perspectives (eds Skouby, K. E., Williams, I. & Gyamfi, A.) 191–210 (River, 2017).

Gebre, G. G., Isoda, H., Rahut, D. B., Amekawa, Y. & Nomura, H. GeoJournal 86 , 843–864 (2021).

Mintah, K., Boateng, F. G., Baako, K. T., Gaisie, E. & Otcherea, G. K. Land Use Policy 109 , 105635 (2021).

Food & Agriculture Organization. The White/Wiphala Paper on Indigenous Peoples’ Food Systems (FAO, 2021).

Tribal Adaptation Menu Team. Dibaginjigaadeg Anishinaabe Ezhitwaad: A Tribal Climate Adaptation Menu (Great Lakes Indian Fish and Wildlife Commission, 2019).

Lubchenco, J., Haugan, P. M. & Pangestu, M. E. Nature 588 , 30–32 (2020).

Article PubMed Google Scholar

Download references

Reprints and permissions