Individual factors affecting the nutrition security of households in the urban areas of Iran

Introduction: The goal of a food and nutrition (FN) system is to provide continuously everybody with adequate healthy food consistent with their preferences. Failure to achieve this goal is a major obstacle to attain individual and social development goals. Then, it is a matter of importance to examine the FN status of households and identify the factors affecting their FN status. This research is conducted to identify the individual factors (especially size of the household, household head education and his marital status) affecting the nutrition security of households in urban areas of Iran. Food security is defined as the availability and the access of food to all people, whereas nutrition security demands the intake of a wide range of foods which provides the essential needed nutrients.
Methodology: We analyzed the data from a sample of 18809 households living in all the provinces in Iran in 2016. The data were collected by the Statistical Center of Iran through Household Expenditure and Income project. A nutrition insecurity index was defined based on the “suitable and preferable food basket for Iranian households”. There are 14 food groups required to meet essential micronutrient needs for human body. For each group a minimum standard level is determined. The household is moderately insecure if it is undernourished in more than seven food groups and severely insecure if it is undernourished in more than 10 food groups. The insecure status of a household is denoted by 1 and its secure status is marked by 0. As the regressor takes two values (0 or 1), the Logit model is employed to run the empirical model.
Regressors are a variety of socio-economic variables. Through trial and error, we found seven factors determining the nutrition security of the households including annual per capita income of the household, house ownership status, age of household head, contribution ratio (ratio of the number of employed members to household size), marital status of the household head, the household head’s education, and household size (the number of members). Multicollinearity was also examined. In a logit model, F-statistic test is not suitable to identify multicollinearity. Therefore, we used partial correlation coefficients and VIF tests. 
Results and Discussion: The results showed that 82 percent of the households in more than seven groups (moderate nutrition insecurity) and 45 percent of the households in more than 10 groups (severe nutrition insecurity) out of 14 food groups were undernourished.
In case of moderate nutrition insecurity, all of the explanatory variables were significant at the level of 1 percent p-value, but, for household head education, it was significant at 10 percent. Also, all the variables explained the severe nutrition insecurity at the level of 1 percent.
The marginal effects of the determining factors were also calculated. The annual per capita income, the contribution rate and the size of the household were known as the most effective determinants of nutrition security of the household. The marital status of the household head and house ownership came next. Household age and education were slightly effective as well. The households with married heads were more nutrition secure although their annual per capita expenditure on food was less than those with single heads. Marriage seems to have a positive effect on the household food standards and quality.
Few researches have met the fact that education has negative effect on nutrition security. This research is one of them. The households in which the heads are of higher education level, in our sample, had higher annual per capita incomes, higher annual per capita food expenditures, bigger sizes and smaller contribution ratios. It seems that although higher education level increases the quantity of food consumption, it fails to make the quality of it better..
All the partial correlation coefficients of the explanatory variables were significant at 5 percent (except for contribution ratio and house ownership, which was insignificant). But the biggest coefficient was 45 percent, so small that we could say there was no sign of multicollinearity. Of course, it does not confirm the lack of multicollinearity. VIF also did not show any clue of multicollinearity. All the variable VIFs were between 1 and 2, much smaller than 10, which is alarming.
Conclusion: In search of the factors affecting the nutrition security of urban households in Iran, we raised questions about three variables including household size, household head’s marital status and his education level. Our data from 18809 households in all the provinces suggested the following:

As the household size grows bigger, the family becomes more nutrition insecure.
As the household head’s education level is higher, the family becomes more nutrition insecure.
Households with married heads are more nutrition secure. It seems that the nutrition standard and quality is better in these families although they spend less per capita food than households with single heads.

The other factors identified were the annual per capita income of household, house ownership status, age of household head, and contribution ratio.