The dynamic nexus between climate changes, agricultural sustainability and food-water poverty in a panel of selected MENA countries

This study attempts to examine the dynamic relationships between climate changes, agricultural sustainability and food-water poverty in a panel of MENA countries over the period 1990–2016. A panel co-integration, pooled least squares regression, pooled fixed effects, and pooled random effects models with the Hausman test for model specification are used to relate three proxies for food poverty and two proxies for water poverty to standard weather variables, agriculture productivity indicators, and environmental sustainability variables. The main results of regression analysis indicate that out of the three food poverty models, two food poverty regressions indicate the low agricultural productivity in lowand middle-income countries, while water poverty in terms of access to improved water is found to increase substantially agricultural value added (coefficient is more elastic, i.e. more than the unity). The results further show that high precipitation and temperature, often accompanied by high CO2 emissions, increase food poverty in terms of food deficit and prevalence of undernourishment, whilst having no significant effect on water poverty. The overall findings conclude that there is a substantial requirement to increase agricultural sustainability in lowand middle-income MENA countries without deteriorating environment and water reserves. This is an Open Access article distributed under the terms of the Creative Commons Attribution Licence (CC BY 4.0), which permits copying, adaptation and redistribution, provided the original work is properly cited (http://creativecommons.org/licenses/by/4.0/). doi: 10.2166/wcc.2019.309 ://iwaponline.com/jwcc/article-pdf/12/1/1/851488/jwc0120001.pdf Hatem Jemmali (corresponding author) Higher Institute of Accountancy and Administration of Enterprises, University of Manouba, Manouba, Tunisia and Laboratory for Research in Quantitative Development (LAREQUAD), College of Economic Sciences and Management of Tunis, University of Tunis ElManar, Tunis, Tunisia E-mail: hatemjemmali79@gmail.com Rabeh Morrar An-Najah National University, Nablus, West Bank, Palestine Mohamed Safouane Ben Aissa Laboratory for Research in Quantitative Development Economics (LAREQUAD), College of Economic Sciences and Management of Tunis, University of Tunis ElManar, Tunis, Tunisia


INTRODUCTION
The Middle East and North Africa (MENA) is a large, complex, and diverse region, which faces a wide range of economic and food-water security challenges. They are generally classified into two groups; high-income countries, whose economies are dominated by petroleum and natural gas, and low-and middle-income countries with more diversified economic structures and a large contribution of the agricultural sector to the local economics (Pelzman ).
Known to be one of the most arid regions in the world, the MENA region is, furthermore, vulnerable to climate-induced impacts on water resources and food production.
However, despite this alarming situation, political leadership in the region has low awareness about the need to promote adaptive governance strategies to deal with the increased hydrological risk. It becomes clear that adaptation to climate change and environmental sustainability are closely linked to water availability and food production added to a set of social, economic and political factors. In most of the MENA countries, climate change is associated with reduced average precipitation, which creates a real threat to food security in this dry region. In some cases, an intensification of the water cycle has caused more extreme floods and droughts in the region. Generally, climate change acts as a 'threat multiplier' which increases the vulnerabilities of already vulnerable and poor populations, creates threats to security, and raises risks of violence in the region.
In the MENA region, the linkage between food-water poverty, agricultural sustainability, and environment conservation has not received enough policy attraction in favor of the most vulnerable groups who are most affected by the climate change, food security challenges and limited accessibility to water resources. However, rich MENA countries (oil-based countries) have been successful in tackling food-water poverty, at least in the short term, which was made possible by using oil rents.
The idea for an integrated food-water nexus was mainly highlighted in the agenda of the Bonn 2011 conference about the sustainable security systems framework (Leese & Meisch ). However before, Blake () emphasized the need to increase food production to fulfill the food requirements of Asia's growing population. His study concluded with policy strategies to attain agricultural sustainability in the region. Schaller () presented the concept of agricultural sustainability as a viable instrument for: (i) sound environmental policies, (ii) amplified economic growth, and (iii) productive rural development, which is responsible for global food production. Heller & Keoleian () explained the long-term sustainability of the US food system to the changes in consumption behavior across agricultural production, distribution, and food disposition. Zezza & Tasciotti () examined the relationship between urban agriculture, food security, and poverty issues using household survey data of 15 developing countries. They found that the contribution of the agricultural sector to GDP is very low, which underestimates the value added of urban agriculture and its ability to reduce food insecurity and urban poverty. Kemmler & Spreng () confirmed the alignment between human activities and sustainability issues. They added that 'the energy system is a sound framework for providing lead indicators for sustainable development'. Ozturk () investigated the dynamic nexus between agricultural sustainability and food-energy-water poverty in a panel of selected sub-Saharan African countries over the period 1980-2013. He confirmed the importance of water resources for a better quality of life and food challenges, which is deemed desirable for water management mainly in the agricultural sector. He recommended developing sound institutions and technological upgrades to fulfill the energy-water food nexus for sustainable development (see also Kaygusuz () and Rasul ()).
This study is expected to contribute to the existing literature by exploring the main determinants of food-water poverty which are associated with the agricultural growth factors, environmental sustainability, and climate change. This is crucial for making policies which are relevant to the long-term agricultural development and ensuring food and water security in the region. The rest of the paper is organized as follows: the next section introduces a brief overview of the MENA region, and why it is pressing to focus on matters like food-water poverty, agricultural sustainability, environmental degradation and climate change in the region. The following section discusses the main data sources and the methodology for empirical analysis.
The results and discussion follow, and the last section consists of the conclusions.

THE MENA REGION: AN OVERVIEW
One of the main peculiarities of the MENA countries is their heterogeneity in economic, social, and political situations and environmental conditions. They can be divided into two main groups based on their gross national income (GNI) (see Table 1). The first group includes the least populated and high-income economies, which are mainly based on petroleum and gas rents, and with low contribution of the agriculture sector to employment, exports and GNI due to their severe climate conditions (see Figures A1 and A2, Appendix A). The second group encompasses the low-and middle-income economies, where population growth brings additional challenges.
GNI for this group is quite low and the agriculture sector is flourishing despite the insufficiency of water resources (quality and quantity) (see Figures A1 and A2).
According to the Food and Agriculture Organization's (FAO) statistics in 2016, the proportion of cultivated area for the MENA countries remained low at 11.34%, which is below the world average of 18.33% (see Figure A3) (FAO ). This alarming situation is mostly attributed to the inadequacy of water resources for agricultural products.
Most of the MENA counties suffer from limited water resources and severe deterioration in water quality, which are attributed to the galloping demographic growth that causes over-extraction of groundwater resources and their contamination.
The hydric situation of the MENA region is also aggravated by the climate-induced influences coupled with steady demographic growth which call for urgent plans and adaptive governance strategies to deal with the increasing hydrological risks (Jemmali & Sullivan ).
Acceleration in the hydrological cycle is highly correlated with the fluctuation, disruption and irregularity of rainfall, which sometimes brings flooding and desertification.
The water crises in the MENA region are expected to continue and reach critical levels in the long run (see   Water availability and food security are tightly coupled. The agriculture sector is the highest consumer of water in the MENA region; it relies on both public or private irrigation sources using surface water, groundwater, or both. On average, the agriculture sector accounts for approximately 80% of water expenditure in the region (World Bank ). This means that any plan or action to reallocate water under conditions of scarcity will most likely be at the expense of food security, which has been experienced in Yemen, Jordan, Israel, and Libya (World Bank ).
Also, climate change is expected to have a significant impact on food production (Cline ). Egyptian agronomists have estimated the impact of climate change on food supply using a combination of standard global circulation climate models with multi-year and multi-crop models.
They found that water demand increased for most crops (grains, including maize, wheat, sorghum, barley, and rice) due to higher temperatures and lower yields. Also, all crops experienced a significant decline in production, ranging from 9 to 19% for a 2 C average temperature rise, along with increased water consumption of 2-16% (Eid Finally, it is worth noting that food insecurity has negative consequences on poverty rates and undernourishment, which are more apparent in rural areas in Iraq, Sudan and Yemen (Lofgren & Richards ) (see Figure A5). Studies expect that MENA's population will continue to increase sharply until 2050, whereas the current food production system is expected to face deficiencies in the provision of the region's needs from foods by 2050.

DATA AND METHODOLOGY
The data on MENA countries (the study follows the definition of the MENA region adopted by the World Bank which includes Israel, Iran, Malta and the Arab world, minus Mauritania) for food-water poverty, agricultural sustainability, environment and climate changes (see Table 1) between 1990 and 2016 were obtained mainly from the recent World Development Indicators published by the World Bank () and the International Financial Statistics published by the IMF (). Data on rainfall and temperature were taken from the recent statistics of the NOAA (National Oceanic and Atmospheric Administration). In order to avoid omitting some countries due to missing values, the forward and backward interpolation technique was used to fill these gaps.
The empirical model includes five dependent variables (response variables), including three food poverty indicators and two water poverty indicators that were separately regressed with a set of explanatory variables (see Appendix C). It is worth noting in this regard that foodwater poverty is a buzzword that is widely used by the policymakers to assess the inadequate intake of food calories per day and inadequate access to water resources among households across countries. In order to deeply comprehend the food-water poverty nexus, one may look to the definitions of these concepts (see Appendix B). The used variables were chosen to give broader coverage of food-water poverty and agricultural and environmental sustainability in the region. The descriptive analysis will be followed by both pooled and panel regressions. Following Ozturk (), to understand more deeply the food-water poverty, agricultural sustainability and climate change inter-linkages, we used the following models.
Model A: Food poverty where WaPOVk represents the water poverty indicators limited to: wapov1 (percentage of population without access to water sources) and wapov2 (percentage of population without access to sanitation facilities).
Here After taking the logarithm form, the fixed effect models to estimate can be written as below: where α i is a country specific effect.
Subsequently, in order to absorb the MENA countries' specific shocks, the study used a panel random effects model that could take the following forms: where β t represents time variant shocks.
It is noteworthy that the classical Hausman () test could not be used to decide whether the fixed effects regression is better than the random effects or vice versa as the estimates of fixed and random models are all robust.

RESULTS AND DISCUSSION
A descriptive analysis is performed to show the basic characteristics of data before delving into the main question of the study which discusses the main determinants of food-water poverty in the MENA region. Figure 1    To find out if there is a co-integration or long-run association between the variables, we use two known tests: the Johansen Fisher Panel co-integration test and the Kao residual co-integration test based on the null hypothesis of no co-integration against the alternative hypothesis of the co-integration equation among the variables. The results in Table 4 reject the null hypothesis of no co-integration, while it accepts the alternative hypothesis of a co-integration relationship in the five food-water poverty models. According to the Fisher-type Johansen test, the co-integrating equations that exceed four equations imply the presence of The pooled least squares regression method is used first to obtain the estimates for each model. Table 5   between food challenges and agricultural and environmental sustainability and economic growth.
Another interesting result in Table 5 shows that an increase in the food deficit leads to an increase in carbon dioxide emissions and fossil fuel energy consumption by 0.540 and 9.393%, respectively, which implies that environmental degradation is strongly associated with the depth in food poverty. This requires cleaner production techniques in the region to transform food production into a sustainable mode (see Lebel & Lorek ; Cohen ; Bogdahn ).
In other words, the need for sustainable consumption and production by renewable, clean and efficient energy sources is strongly recommended.
Results for the variables rainfall and lack of access to water in model 1 are surprising. When precipitation and access to adequate water increase, the depth of the food deficit also increases with positive coefficients of 0.4 and 0.377 respectively. This may be explained by rainfall variability. It is known that variability of high-frequency precipitation may reduce land suitability and crop yields and accordingly may have a negative effect on food production.  From the food poverty models, it is easy to conclude that rainfall and temperature variability added to lack of access to improved water can have adverse effects on the viability of economic systems, food production, and food availability.
They may increase the percentage of total undernourished population and food deficit depth. Agricultural sustainability is shown, therefore, to play a key role in combating food poverty issues.
Turning to water poverty models, models 4 and 5 in Table 5 reveal that an increase in temperature leads to more water poverty in regards to limited accessibility to improved water (5.410%) and less water poverty in regards to more accessibility to sanitation facilities (À4.544%).
Temperature variability may offset the water resources available for drinking. Also, the prevalence of undernourishment is negatively associated with cereal yields (À1.36%) and with CO 2 emissions (3.407%), and sanitation poverty is negatively associated with cereal yields (À1.336%) and CO 2 emissions (À5.86%). In this regard, previous literature suggested a set of sustainability options to improve agricultural productivity and ensure water security under climatic variations. Table 6 shows the estimates of different food-water poverty models using the fixed effect regression method. The results show that climate changes denoted by temperature have no significant impact on food-water poverty models, except for the FoodPOV3 model which exhibits a significant and negative relationship between temperature and the The results shown in Table 6 relating to water poverty in terms of access to improved water exhibit a significant, positive and more elastic influence with agriculture value Note: Robust standard errors in parentheses, ***p < 0.01, **p < 0.05, *p < 0.1.
All the variables are in natural log form. Finally, Table 7 shows the estimates of pooled random effects regression for the food-poverty model. The weather factors (rainfall and temperature), which are widely considered as the main indicators of climate change, are found to have a positive effect on the prevalence of undernourishment similarly to the OLS results. As explained above, the positive effect of rainfall on food deficit (0.370%) and prevalence of undernourishment (0.277%) Note: Robust standard errors in parentheses, ***p < 0.01, **p < 0.05, *p < 0.1.
All the variables are in natural log form. may be explained by the rainfall variability which may reduce land suitability and crop yields and accordingly have a negative influence on food production. The results show further that temperature has a negative effect on household income (À0.292%) and a positive (elasticity more than 1) relationship with the food prevalence of undernourishment. This could be explained by the fact that an increase in temperature may reduce seasonal agriculture which is critical for food security in many low-and medium-income countries, as stated in many previous studies.
The first and third columns of

CONCLUSIONS
In this study, the food-water poverty nexus was investigated against climate changes and agriculture sustainability. The current study used the depth of the food deficit, final household consumption expenditure per capita, and the prevalence of undernourishment as three proxies for food poverty, and the percentage of the population without access to a safe water sources and percentage of the population without access to sanitation facilities as two proxies for water poverty.
These proxies served in the empirical analysis as the dependent variables for different food-water models. In addition, this study used as independent variables two climate variables, namely annual average of temperature and rainfall, and some agricultural and environmental sustainability indicators, i.e. agricultural value added, cereal yields, forest area, carbon dioxide emissions, and fossil fuel energy consumption. A few other variables were also added to different models to take into account economic welfare and the global level of prices, including GDP per capita and inflation.
We found some limitations in using the same food-poverty proxies mentioned in the literature. For instance, we may use the Water Poverty Index, widely used to assess the multidimensional water poverty at different scales (Jemmali & Sullivan ) in the regression analysis to replace the simple water access indicators. Also, other food security indices (e.g. the Global Food Security Indexindex developed by The Economist Group) may be used instead of the aforementioned food indicators. One of the main advantages of using such food security indices in a dynamic (quantitative and qualitative) benchmarking model is the ability to assess the exposure of a country to the impacts of a changing climate, its susceptibility to natural resource risks, and how it can adapt to these risks.
The main findings in different regression approaches revealed that out of the three food poverty indicators, two of them indicated the low agricultural productivity among lowand middle-income countries in the region. Regarding the impact of climate changes on the food-poverty nexus, we found that precipitation and temperature positively affect food poverty in terms of food deficit and prevalence of undernourishment, but not for water poverty indicators. The increase and high variability of rainfall in some countries added to the remarkable increase in temperature may contribute to reducing food production and cause high levels of food poverty. Agriculture value added was negatively associated with three food poverty indicators, and positively associated with the lack of access to improved water. Also, the results exhibited that CO 2 emission positively influenced food deficit and prevalence of undernourishment, which means that an increase of food poverty could be accompanied by environmental degradation and high levels of air pollution in some MENA countries. Fossil fuel energy consumption and economic growth were found to have different impacts on the aforementioned food poverty indicators. In an unexpected result, we found that the lack of access to improved water may increase substantially with an increase in agricultural value added, while the economic growth tends to show a negative association with water poverty.
Thus, to alleviate food-water poverty levels, it is crucial to prompt agricultural productivity without damaging the environment. This dilemma of food-water poverty and agricultural sustainability should be a priority on the agenda of policymakers in the MENA region. There is an increasing need to formulate appropriate policies that align between food security for poor people and regular access to improved water and sanitation facilities. It is found that less developed countries in the region are the most exposed to climate changes and severe water and food insecurity. In order to mitigate the unpredictable impacts of high rainfall variability and increasing temperature, more sustainable and efficient agricultural practices should be implemented.

SUPPLEMENTARY MATERIAL
The Supplementary Material for this paper is available online at http://dx.doi.org/10.2166/wcc.2019.309.