Contents - Previous - Next
This is the old United Nations University website. Visit the new site at http://unu.edu
Determinants of family and preschooler food consumption
International Food Policy Research Institute, Washington, D.C., USA
The assumption in many previous nutrition interventions has been that increases in food expenditures improve family caloric intake and ultimately increase child energy consumption and improve a child's nutritional status. An intervention that has a positive, significant effect on any one of these linkages is presumed ultimately to improve child health and nutritional status. For example, food stamp programmes assume that by increasing a family's food purchasing power, the family diet will improve and a part of this nutritional benefit will be passed on to the child. Similarly, a supplementary feeding scheme is based on the assumption that provision of food-either to a specific individual or to a family-will ultimately lead to an improvement in the recipient's nutritional status.
Existing data sets from four countries were used for the analyses in this paper. The Thai data were collected as part of a nation-wide survey (1). The Mexican data refer only to Mexico City and were collected in May 1978 as part of an evaluation of the CONASUPO milk subsidy programme (2). Similarly, the Sudan data were collected only for an urban population (Khartoum) as part of an evaluation of the government's wheat subsidy programme (3). The Malaysia data were generated from an evaluation of the World Bank irrigation project in the Muda Region (4).
Specifically, this paper is concerned with three questions:
EFFECT OF INCOME ON CALORIC INTAKE
The first analyses were directed toward the issue of the effect of income on the caloric intake of a population. Data from Thailand were used to examine the impact of increasing income on caloric intake of both urban and rural groups. Malaysian expenditure data for a rural sample and Sudan income data for an urban population were also analysed. As can be seen from tables 1 to 3, as income increases the mean caloric intake within a population also increases. This is true for both rural and urban areas. However, in Thailand, at each level of income, urban groups show consistently lower caloric intakes per capita than do rural groups in that country (table 1). This could reflect one of two situations. First, the data could be interpreted to mean that energy intakes as a proportion of requirements at any given level of income for urban groups are lower than for their rural counterparts. However, an equally plausible explanation is that activity levels are lower in urban areas and thus urban populations have adjusted their caloric intake downward to reflect their decreased energy needs. Without data on actual activity patterns, it is difficult to substantiate this latter hypothesis.
Because family income is often difficult to measure precisely in large-scale surveys, we used total monthly expenditures as a proxy for income to determine if the pattern of caloric intake per capita was different from that shown in tables 1 to 3. For Thailand (table 4) and the Sudan (table 5), as total expenditures increase there is a concurrent increase in caloric intake. Here again for Thailand, rural households at each level of expenditure have consistently higher per capita caloric intakes than do the urban groups.
TABLE 1. Mean Caloric Intake per Capita by Decile of Monthly Income-Thailand, Urban and Rural
TABLE 2. Mean Caloric Intake per Capita by Decile of Monthly Income-Khartoum, Sudan
* Figures in parentheses = N.
** Millet, corn, rice, and macaroni,
***Weighted average (families consuming one cereal may not be consuming others).
TABLE 3. Calories Purchased per Capita by Decile of Expenditure Group*-Malaysia
|Decile||X kcal Intake||N|
* Total expenditures used as a proxy for income.
TABLE 4. Mean Caloric Intake per Capita by Decile of Monthly Total Expenditure-Thailand, Urban and Rural
TABLE 5. Mean Caloric Intake per Capita by Decile of Monthly Total Expenditures-the Sudan
N = 489 in each docile.
TABLE 6. Mean Caloric Intake per Capita by Decile of Monthly Total Expenditures-Urban Thailand
TABLE 7. Mean Caloric Intake per Capita by Decile of Monthly Total Expenditures-Rural Thailand
TABLE 8. Caloric Intake per Capita in Each Major Food Group by Decile of Total Expenditures-Muda Region. Malaysia
|Polished and glutinous rice||1,257||1,259||1,196||1,215||1,332||1,251||1,436||1,452||1,454||1,474|
|Potatoes and other roots||11||12||13||13||19||14||18||16||17||18|
|Fresh local fruit||19||22||23||30||33||32||41||44||58||63|
TABLE 9. Cross-classification of Households on the Basis of Income Level and Caloric Intake-Thailand, Urban and Rural
Income level *
Income level *
|< B 291.14||> B 291.14||< B 205.15||> B 205.15|
|< 2,500 kcal||313||1,221||1,534||1,591||1,841||3,432|
|> 2,500 kcal||186||1,226||1,412||1,085||2,618||3,703|
* The division of income levels is based on the urban poverty
line of 291.14 baht and the rural poverty line of 205.15 baht.
Source: Trairatvorakul (5).
In general, for these three countries-Thailand, Malaysia, and the Sudan-caloric intake increases as one moves from the lowest to the highest income and/or expenditure group.
We were also interested in the effect of income on the amounts and types of food consumed. Tables 6 to 8 show the relative contribution of major food groups to caloric intake, broken down by expenditure group in Thailand and Malaysia. The results are similar for the two countries. The relative contribution of calories from the various food groups is fairly stable across each of the ten income groups. In Thailand there is some shifting of calories between glutinous and non glutinous rice. What these data suggest is that in these countries, as income increases, people are consuming more of the same diet rather than changing to a different mix of foods.
If the data were analysed simply this way, the results would suggest that low income is a good indicator of households that are likely to have inadequate caloric intakes. However, a cross-classification of households using income levels and caloric intake (table 9) shows that some high-income households, both urban and rural, have inadequate caloric intakes (5).
These data suggest that the income criterion does not identify all households with an inadequate food intake. Some high-income families appear to consume too few calories. Conversely, not ail poverty level households have insufficient caloric intakes; in rural Thailand, approximately 40 per cent of all the low-income families consume a nutritionally adequate diet. But why? The following section attempts to identify the key determinants of family food consumption.
DETERMINANTS OF FAMILY AND PRESCHOOLER CALORIC INTAKE
From the previous analyses we have seen that caloric intake increases as income increases; the magnitude of this change may differ by urban/rural location and may also differ by country. While this income/caloric relationship is of interest, what is of more interest to many policy makers is the effect of increased family income on the dietary status of the household and of individuals within the household. We were interested in the effect of an increase in income on the family's caloric intake as well as the energy intake of preschoolers within the household.
In order to explore the income/family consumption/ preschooler nutritional status linkages, two data sets were used-from Malaysia and from Mexico. The results of the analyses for each country will be presented first, followed by a discussion of the results.
Anthropometric measurements-weight, age, and length or height-were used to assess preschooler nutritional status in each of the data sets. In order to be able to aggregate each of the anthropometric measurements across age groups from birth to five years, a Z-score for weight for age and height for age was computed for each observation. The 1976 National Center for Health Statistics Growth Standards (6) were used as the standards for weight and height. The mean Z-scores for weight/age and height/age were .3664 and-.4336 for Malaysia and -.1474 and -.6178 for Mexico. The more negative the Z-score, the more it deviates from the median standard for that age. Thus, the more negative a particular Z-score is, the more malnourished, on average, is the group of children. What is apparent from these aggregate scores is that the problem in these areas is one of chronic malnutrition as evidenced by the low height/age scores across the two countries.
Initially we were interested in the effect of family level variables, in particular, household incomes, on the nutritional status of the child. Tables 10 and 11 show that for both Malaysia and Mexico household income* is not significantly associated with any of the Z-scores. For both countries, family level factors explain little of the variance in preschooler nutritional status-either weight or height. The Rē for each of these equations is amazingly low, and therefore the equations have no utility in predicting preschooler nutritional status.
An additional analysis exploring the income/family caloric intake relationship was done for Mexico. Family energy intake was measured using a 24-hour recall of total household food consumption, including food eaten away from home. Family caloric intake is positively and significantly affected by increments in household income (table 12). Interestingly, participation in the milk subsidy programme in Mexico City also significantly increased the number of calories consumed within the family; programme participation was associated with an approximately 1,700 calorie increase in family energy intake. The family benefited as a result of participation in the milk subsidy programme.
TABLE 10. Z-Score Weight and Height Regressed on Family Level Variables-Malaysia
|Total yearly family expenditures*||-.13-05||.57-06|
|Years in project||-.94-02||.17|
|Group (participation/||.32||- .57|
N = 424 t-statistics in parentheses
* Proxy for income
** Significant at 5% level
TABLE 11. Z-Score Weight and Height Regressed on Family-Level Variables-Mexico
|Weekly family income||.39-04||.47-04|
t-statistics in parentheses
TABLE 12. Regression for Daily Family Caloric Intake-Mexico
|Weekly family income||.399|
|Family water supply||974|
t-statistics in parentheses
* Significant at 5% level
** Significant at 1% level
Having determined from the Mexican data that increased family income improved family caloric intake, we were interested in the effect of income on the preschoolers' calorie consumption. Table 13 shows the mean caloric intake for preschoolers participating and not participating in the Mexican milk subsidy programme; both groups have average energy intakes below the recommended level. Milk contributed significantly more to caloric and protein intake of the subsidized children than of the children from non-subsidized households.
TABLE 13. Mean Caloric Intakes for Programme and Non-programme Children-Mexico
of Milk to
of Milk to
|Non-milk programme preschoolers||92.2||19.3||25.5|
TABLE 14. Regression of Child Caloric Intake on Family-Level Variables-Mexico
|Age (in months)||2.67||Age||3.05|
|Weekly family income||-.26 - 01||Family size||-66.7|
|Family size||15.4||Number of children under six years||-35.0|
|Sex (0 = male, 1 = female)||- 26.0||Sex (0 = male, 1 = female)||42.0|
|Number of children
under six years
|-96||Family caloric intake||.42 - 01|
|Group (participation/||34||Group (participation/||62|
t-statistics in parentheses
* Significant at 1% level
However, table 14 shows that the preschoolers' calorie intake is not significantly associated with weekly family income. In addition, unlike what we observed for family energy intake, the child's caloric consumption was not significantly increased by participation in the milk subsidy programme. So, although the programme children consumed more milk, their caloric intake was not significantly increased as a result of the milk subsidy scheme.
The data from table 14 also indicate that family caloric intake has a positive, significant effect on the preschoolers' energy intake.
DISCUSSION AND COMMENT
The results given in the section on the effect of income on caloric intake indicate that, in general, as income increases, caloric consumption also increases. However, the Thailand data (table 9) also indicate that there are high-income households who fail to achieve caloric adequacy. The reasons for this are unclear. Clearly, income is only one determinant of family caloric intake.
The Mexican results show that as income increases, the caloric intake of the family is improved. All else being equal, higher-income families are more likely to have an adequate dietary status than low-income households. However, in Mexico, increased income is associated with an increased energy intake in the child only to the extent that the income is used to purchase additional family calories. There was no direct effect of income on preschool nutritional status as judged by weight and height for age for either Mexico or Malaysia. This finding is consistent with other recent studies. For example, a study (7) in Nicaragua found that income was not a significant factor in explaining child anthropometric status.
We know that aggregate supply of food or nutrients within a country is not a sensitive measure of household food consumption. It now appears from these data that caloric availability within the household may not be a precise indicator of a child's nutritional status. It is only to the degree that children receive a portion of the incremental calories within the family that preschooler energy intake will improve.
The Mexico milk subsidy did increase the caloric intake of the household by 1,708 calories, or the equivalent of 248 calories per individual household member. Why did this 248 calories not make a significant net addition to the diet of the preschool child? There are two plausible explanations. First, the malnutrition problem in the Mexico study sample is predominantly mild/moderate chronic malnutrition. It is not the type of wasting and stunting that would be visually diagnosed by the family. It may very likely be that there is not the perception on the part of household members that the preschool child needs more calories. If all children in the community are stunted, the programme child looks just like every other preschooler.
Second, there may be a lack of demand for food on the part of the child. If the child appears satiated-whether because he is satisfied or because of a general anorexia- there would be no reason for the family to assume the child needs or wants more food. This last point has been given little attention in earlier studies and needs to be explored in future work.
What the present analyses suggest is that higher income is not necessarily sufficient to ensure adequate caloric intake within the household or by individual family members. In addition, family-oriented nutrition intervention may not be the most effective means of achieving nutritional goals for the specific household member. Future work on this latter point is needed.
1. Government of Thailand, National Statistical Office, National Socio-economic Survey of 1975-1976 (Bangkok, 1977).
2. C. Overholt et al., "Case Study: Subsidized Milk Distribution in Mexico," ch. 5 in J. E. Austin and M. Zeitlin, eds., Consumer Food Price Subsidies (Oelgeschlager, Gunn and Hain, Cambridge, Mass., USA, 19811.
3. Government of Sudan, Department of Statistics, 1978/1979 Household Budget Survey ( Khartoum, 1979).
4. Food and Agriculture Organization of the United Nations and World Bank, "1974 Muda Farm Household Survey" (World Bank, Washington, D.C.).
5. P. Trairatvorakul, "Rice Price and Calorie Intake of the Thais (International Food Policy Research Institute, Washington, D.C., 1983).
6. US Department of Health, Education, and Welfare, National Center for Health Statistics, "NCHS Growth Curves for Children, Birth-18 Years" (Washington, D.C., 1977).
7. Barbara C. Wolfe and Jere K. Behrman, "Determinants of Child Mortality: Health and Nutrition in Developing Countries," Econ. Dev., 11:63 (1982).
Contents - Previous - Next