AN EMPIRICAL STUDY ON THE IMPACT OF ECONOMIC DEVELOPMENT IN FOOD SECURITY IN EMERGING AND DEVELOPED ECONOMIES

subdivided


Analysis of the evolution of food security
We consider models that include a linear trend and quadratic 1 trend specific to each country. The first model is written as follows: With:

PREVAL :
The variable relating to the prevalence of undernourishment.
TREND : Linear trend.The quadratic trend model is written as follows: Quadratic linear trend.
The coefficient for the quadratic term (  ),gives information about the shape of the curve. The coefficient (α) indicates the direction of the curve. If (α) is positive then we can conclude that malnutrition increases over time.   Food security and economic growth Along these lines, we proceed in a dynamic panel data model. Indeed, this model is more powerful than static since it allows to eliminate the term specific individual heterogeneity ( i  ) and offers, therefore, a better efficiency of the estimators. The method used is that of GMM system of Arellano and Bond (1991). The command used is Xtabond2, under the STATA software 12.0 . We will, first, estimate a model for the whole sample (26 countries). Indeed, the basic equation can be described as follows:  The dependent variables: , it PREVAL : The population below minimum level of dietary energy consumption (also referred to the prevalence of undernourishment) shows the percentage of the population whose food intake is insufficient to meet dietary energy requirements continuously.
, it DEFICIT : The extent of the food gap is the amount of calories that lack an undernourished population to no longer be considered as such, all things being otherwise equal. The average intensity of food deprivation of undernourished people, which corresponds to the difference between the average dietary energy requirements and the average dietary energy intake of undernourished population is multiplied by the number of undernourished people to get estimate of the total existing food deficit in the country, a figure which is then adjusted to the total population. (Source :http://www.fao.org/docrep/019/as212f/as212f.pdf) The explanatory variables: , it PIBtete : It is the GDP per capita, it is a proxy for economic growth.
Sante : Share of health expenditure in GDP. Total health expenditure is the sum of public and private health expenditures. It covers the provision of health services (preventive and curative), family planning activities, activities related to nutrition and using emergency reserved to health but excludes the provision of water and hygiene services.
, it alpha :Adult literacy rate (15 years +) (%). It is the percentage of the population aged 15 and over who can understand, read and write short statements about her daily life. Generally, literacy also includes numeracy, that is to say the ability to perform simple arithmetic operations. This indicator is calculated by dividing the number of literates aged 15 and over by the population of the relevant age group and multiplying the result by 100.
, it improvedwater :Access to improved water source is the percentage of the population with reasonable access to an adequate amount of water coming from an improved source such as a household outlet of water, public standpipe, a well, a spring or a protected well or collected rainwater. Unimproved sources include vendors, tanker trucks and unprotected wells and springs. Reasonable access is defined as the availability of at least 20 liters per person from a source located within one kilometer of the dwelling.  Source: own elaboration based on the outputs of Eviews 9.0 software. -All Variables do not follow a normal distribution. Indeed, the P-Value related to Jarque-Beratest is less than 5%.
-The group number 3 has the highest growth rate and the lowest prevalence. In contrast, Group 1 has the lowest growth, and group 2 exhibits the highest prevalence rate.
-The Group 1 has the highest share of health spending in GDP compared to group 3 and group 2. The last group has the lowest share.

953
The Correlation Analysis   own elaboration based on the outputs of Eviews 9.0 software -The Correlation matrix shows that the prevalence of undernourishment is negatively correlated with health spending and access to an improved water source, that's for the total sample. -For the group 1, it is characterized by a strong negative correlation between the rate of prevalence and IMPROVEWATER and the literacy rate.
-For Group 2, it is characterized by a strong negative correlation between health spending and the variable to undernourishment.
-With respect to group 3, it is characterized by a negative correlation with the prevalence of access to water, spending on health and literacy rates.      The comparative analysis of the evolution of the economic growth rate and the prevalence of undernourishment highlights the following remarks: -The undernourishment prevalence rate seems to go in the opposite direction with the economic growth rate of the total sample, and that's for the entire period.
-For Group 1, it appears that the rate of economic growth has failed to compromise undernourishment during the period 2006-2003. For group 2, during the period from 1992 to 1994, the rate of economic growth seems to be going in the same direction with the non-fed population rates.
-In regards with the group number 3, throughout the study period, the malnutrition prevalence rate seems countercyclical. .06(0.321) (***), (**) and (*) respectively correspond to the statistical significance of 1%, 5% and 10%. m2 indicate the test of serial correlation of order 2 between residues. The Sargan test of validity means the instrument test (p-value).

Results of estimates and interpretations
Source: own elaboration based on the outputs of the Stata 12.0.   (***), (**) and (*) respectively correspond to the statistical significance of 1%, 5% and 10%. m2 indicate the test of serial correlation of order 2 between residues. The Sargan test of validity means the instrument test (pvalue). Source: own elaboration based on the outputs of the Stata 12.0. (***), (**) and (*) respectively correspond to the statistical significance of 1%, 5% and 10%. m2 indicate the test of serial correlation of order 2 between residues. The Sargan test of validity means the instrument test (p-value). Source: own elaboration based on the outputs of the Stata 12.0.
Through the software STATA 12.0, we got the table presented above translating the results of our panel data estimation in a Dynamic approach.
The first observation concerns the general model specification. Indeed, the specification is not rejected by the test of over-identification of Sargan. We accept, thereby the validity of the instruments used. Similarly, there is the absence of serial correlation of orders 2 of the residues.
As expected, the rate of economic growth, access to water, health care spending and the literacy rate exhibit negative and statistically significant impact on the prevalence rates of undernourishment and food shortages. Improved growth, access to water and literacy rates lead to reductions in the rate of non-fed population. This result is Health care spending is a factor that improves food security in the entire sample. Good quality nutrition reduces the likelihood of the emergence of serious diseases caused by poor diet. Food security requires a combination of adequate dietary intake and a healthy environment. This is true for all three groups, since this variable and negatively impacts significantly under-nutrition and / or food shortages.  The literacy rate has a negative and statistically significant impact on undernourishment. A better education system is suitable for food security. Better education facilitates better knowledge in food production and resource management. In the same vein, the authors showed that equity between men and women has a positive impact on the use and food security. The results show that a better growth rate contributes to the increase in the proportion of the population nourished. Indeed, the coefficient assigned to the variable related to the growth rate is negative and statistically significant, and this for the total sample and for all groups. Certainly the impact of the growth rate was higher in group 3. This indicates that the impact of growth is faster on food security of the remaining groups. This result is consistent with the earlier analysis that showed that this group is characterized by undernourishment deteriorating at a decreasing rate.
Concerning the variables related to regional dummies, it positively affects undernourishment and in a statistically signification. Sub-saharian Africa seems to have the worst food security, known that it displays the highest coefficient.

Conclusion:-
This paper was dedicated to the study of economic growth-food security relationship for a panel of 52 emerging and developing countries. The technique used is the GMM of Arellano and Bond (1991). The estimation results show clearly a negative relationship between economic growth rates and the prevalence of undernourishment. Economic growth in emerging and developing countries seems to be a key factor to reducing poverty and the proportion of the malnourished population but it is'nt the only factor, others factors must be present olso. The rate of economic growth, access to water, health care spending and the literacy rate exhibit an important impact on the prevalence rates of undernourishment and food shortages. Health care spending is a factor that improves food security in the entire sample. Good quality nutrition reduces the likelihood of the emergence of serious diseases caused by poor diet. Food security requires a combination of adequate dietary intake and a healthy environment. The estimator proposed by Arellano and Bond (1991) is based on the first difference of variables Such a transformation deletes the term of heterogeneity ( ).However, a correlation emerges between the dependent variable (

  
).To work around this problem, Arellano and Bond (1991) propose an implementation of the Generalized Moments method. They use instruments for Where  is a parameter vector (  ) and (  ), (W ) is a matrix that contains the lagged dependent variable and the explanatory variables.
The estimator "GMM" in two stages, which is written as follows: Where, (  It is essential to go through a first step that consists of making the appropriate transformation (first difference), and use the matrix ofsuitable instruments ( i Z )and perform a first estimation called "estimation of the first step." The residues of this first estimate will be used ,in a second step, to calculate a matrix (

2-The specification tests: Sargan test (valid instrument Test)
In order that the estimator GMM remains still valid, perform the test of the validity of instruments a Sargan4test. The null hypothesis states that all moments of restrictions for the dynamic specification are met. The test is summarized by the t-statistic that approximately obeys a distribution Chi-two: