Contents - Previous - Next

This is the old United Nations University website. Visit the new site at http://unu.edu

EFFECT OF HOUSEHOLD SIZE ON CALORIE INTAKE

The data in table 19 show a general trend towards a decrease in household size, in terms both of persons and of adult consumption units, as well as an expenditure per capita increase. Thus, per capita expenditure was higher for smaller families. Of course, the increase in per capita expenditure level as we move up expenditure deciles was sharper than that observed while moving up the calorie-level deciles. Consequently, the upper expenditure deciles would have higher expenditures per capita as well as per household.

The larger households spent relatively less on food than the smaller households with comparable total expenditures. Thus, the larger households spent relatively more on expensive food per calorie.

When we examined the possible effect of household composition on the pattern of calorie intake, expenditure, and household size by looking at a relatively composite variable -the ratio of the total number of persons in a household to the number of adults-we found that the ratio was invariant for the most part, but in some instances it seemed to vary inversely with the size of the household.

This observation pertained across calorie deciles, i.e. moving along the rows in table 19 for each expenditure docile.

Across expenditure deciles-moving along the columns in table 19-the household composition, represented by the ratio of persons to adults, did not vary in most cases; for others, especially in the upper calorie deciles, it declined, suggesting a smaller proportion of children in the upper expenditure group for higher calorie intakes.

A more general observation concerning these data can be stated in two parts. By and large there was an inverse relationship between expenditure per capita and the size of a household and a direct relationship between calorie intake and calorie costs, but little relation between size of household and calorie intake. The proportion of total expenditure used for food was nearly invariant with increasing total expenditure.

TABLE 19. Household size and composition

For a given expenditure level, i.e. moving across the rows in table 19, we found that calorie intake and calorie costs were negatively related and that the size of the households declined and the proportion of total expenditure spent on food rose as calorie intake increased.

Comparing cells in the south-east quadrant with those in the north-west quadrant of table 19, we found calorie intake level to be lower even at higher expenditure levels and with relatively smaller household sizes. If we restrict the comparison to the cells for expenditure deciles 6 and 7 and calorie deciles 2 and 3 on the one hand, and expenditure deciles 2 and 3 and calorie deciles 9 and 10 on the other hand, we would find an inverse relationship between per capita expenditure and calorie intake level and a direct relationship between expenditure and calorie costs for households of nearly comparable size. Thus, the general pattern of behaviour of households in regard to nutrition revealed an amalgam of influence of expenditure level, size of households, and taste. In this amalgam, household size and composition had minimal influence.

FACTORS INFLUENCING NUTRITIONAL INTAKE AND FOOD EXPENDITURE

Table 20 gives the magnitude of factors that may have influenced the food expenditure and nutritional intakes of the D and ND households.

We had data for 18 variables, which were classified into two broad categories containing nine variables each: (a) macro-variables or village-level variables and (b) microvariables or household-level variables. There were three dummy variables, one each for the farm labour group, the non-farm occupation group, and the ND group. The last dummy variable was used for regression analysis relating to all households.

TABLE 20. Averages of macro (village level) and micro (household-level) variables by expenditure deciles and by deficient and non-deficient households

a. See Appendix for abbreviations.

TABLE 21. Correlations (zero-order) among variables

Macro-variables (Village-level)

- V₁. Economic variable 1: (a) size of the village in terms of population (no ); (b) ratio of farm labour/cultivators; (c) ratio of non-farm/farm households; (d) irrigated area as percentage of total cropped area; (e) area used for paddy cultivation as percentage of total cropped area; (f) area used for wheat cultivation as percentage of total cropped area; (9) number of cattle per 100 persons; and (h) number of buffalo per 100 persons. V₁ was a production profile with the exception of economic variable (a).
- V₂. Economic variable 2: (a) population/houses (ratio); (b) the ratio of pucca houses/total houses; (c) houses with separate arrangements for bath (percentage of total); (d) houses with tiled roof (percentage of total); and (e) houses with electric lighting (percentage of total). V₂ represented the standard of living as reflected in the level of housing.
- V₃. Environment 1.1: the following components were expressed in terms of distance from the village in kilometres: (a) college; (b) high school; (c) primary school; (d) library. V₃ pertained to educational overheads.
- V₄. Environment 1.2: distance in kilometres from the village: (a) hospital; (b) dispensary or clinic; (c) primary health centre; (d) vaidya physician dispensing herbal medicines); and (e) trained nurse. V4 pertained to health overheads.
- V₅. Environment 2.1: distance in kilometres from the village: (a) railway station; (b) bus stop; (c) post office; (d) telegraph office; (e) telephone office; (f) head quarters of village-level worker; (9) credit co-operatives; and (h) number of radios in the village. The Vs variable represented a broad spectrum of communication, including transportation, information, and extension.
- V₆. Environment 2.2: (a) grocery shop; (b) tea stalls; (c) repair shops (for cycles etc.); (d) ration shop; and (e) number of cloth shops (if any) in village, otherwise distance in kilometres from the village. V6 pertained to consumer market facilities.
- V₇. Education level: (a) percent of literate males; (b) percent of literate females; (c) percent of males with education up to or beyond secondary school.
- V₈;. Social and cultural events from 1974 to 1975:
(a) number of plays; (b) number of cinema shows;
(c) number of cultural programmes. Events were in the village or its vicinity.

To identify the effect of the size of the village we have considered village population separately as V₉.

The macro-variables were evolved in two steps. First, observations for each component were standardized by taking the highest value, making it equal to 100, and expressing observation values for all other villages in terms of an index number which was a percentage of the highest value. Second, the components were combined into a composite macro-variable for each village by using the inverse of the standard deviation for each component as a weight. The macro-variable represented an improved state as the value increased where components were measured in original values. The macro-variable represented deterioration as values increased in cases where original values were measured in kilometres in terms of the distance of the village from the location of a specific agent, institution, or facility. V₁, V₂, V₆, V₇, V₈, and V₉ fell into the first category; V3, V4, and V5 fell mainly into the second category.

Micro-variables (Household-level)

- Total expenditure (rupees per year per capita) (table 10).
- Adults: persons (males, females, children) (table 11) converted into adults, using conversion ratios (table 2).
- H₁. Durable assets score: Index (table 20).
- H₂. Food habit score: index (on lines of durable assets score) (table 2 1).
- H₃. Current borrowings rupees (table 20).
- H₄. Literacy: percentage of literate persons to total persons in the households (table 21).
- H₅. Dependency ratio: ratio of number of non-workers to total workers (table 20).
- H₆. Child ratio: ratio of children to total persons (table 21).

Magnitudes of Variables: Means

Macro-variables: Regarding the economic variables V₁ and V₂, we found that the D households were from better-off villages than their expenditure-decile counterparts. The differences, however, were narrow, except for households in expenditure deciles at the lower or the upper end of the scale. For the V₃, V₄, and V₅ variables representing education, health, and transport facilities, lower values implied easier access to a facility; on the whole, the D households were from villages with easier access to these social facilities. However, the differences were not striking and, as in the case of Vs. the differences could be observed only in the lower deciles.

V₆ represented outlets for consumer articles located in the village. The D households showed a distinct advantage compared to the ND households, since they belonged to villages in which there were, on the average, more retail outlets, including tea stalls.

V₇, when read with the micro variable H₄ (literacy level), suggested that the D households came from villages where literacy was less widespread than in those to which the ND households belonged. However, the D households themselves had a relatively larger percentage of literate members than the ND households. Thus, in villages with lower overall literacy levels the D households had a higher educational status than the ND households, and in villages where a relatively higher literacy level obtained, the ND households had a lower literacy status than the D households.

Via and V₉ did not show a consistent pattern.

Micro-variables: Regarding the durable assets (H₁) and food habits (H₂) variables, the D households had higher scores with one exception in each case. The durable assets represent permanent income position and food habits scores represent preferences for quality of different items of food. H₅, the dependency ratio, showed a mixed pattern: no particular group was consistently in a favoured position. He, i.e. percentage of children to total members, suggested that the D households had fewer children, the difference between the two groups being wider in the lower expenditure deciles. For both groups, the ratio of children to total members declined as per capita expenditure increased: those in upper expenditure deciles had fewer children. The variable for loans, Ha; would imply committed expenditure by way of interest payment and return of loan. The D households had larger loans, but, as total expenditures increased, loans declined. The opposite trend was observed for the ND group of households: for the sixth and seventh deciles the difference was small; for the lower deciles loans were larger for the D households; and for the upper deciles the ND households had larger loans,

Macro-variables and Micro-variables Taken Together

If both macro-variables and micro-variables are considered together, the data suggested that the D households came from villages with distinctly larger numbers of retail outlets for consumer goods, somewhat better housing, a better production profile, and better access to transportation and medical facilities, though not to schools. The villages to which the D households belonged were not larger in size and did not have a better record of entertainment. Households belonging to the D group had a higher literacy rate, a lower percentage of children, a better durable assets position, and a preference for higher-quality foods. For loans and the dependency ratio the pattern was mixed On the whole it seemed that the D households had easy access to consumer goods and preferred a lifestyle in which emphasis on expensive food played a part.

Correlations

The matrix of variables, correlated (a) with expenditure on food, total expenditures, and the squares of the total expenditures, and (b) with individual nutrients, is presented in table 21. We found that, except for carotene, all other nutrients and calorie intake levels were highly correlated. Though adequate intakes of calories or protein or both might not ensure adequate levels of other nutrients, an increase in the former would tend to increase the intake levels of other nutrients.

The size of a household in terms of adult individuals was correlated with nutrient and calorie intake levels. Total expenditures and expenditures on food, both at the household levels, had the highest coefficient of correlation, about 0.9. The magnitude of the correlation coefficients of estimated levels of food expenditures with nutrient intake levels was not so high. Factors other than estimated food expenditure played a minimal role: these were the durable assets score, the food habits score, and the number of adults in a household, followed by the overall literacy rate, and the dependency ratio. The magnitudes of the con relation coefficients were low for these variables.

Part 2 will appear in the next issue.

APPENDIX. DEFINITIONS OF ABBREVIATIONS OF VARIABLES USED IN REGRESSIONS

Dependent Variables

T. Exp. = expenditure per capita per Year (rupees)
Exp. (F), Exp. F. = expenditure per capita per year on food (rupees}
D = calorie-deficient household
ND = calorie-non-deficient household
Exp F. (D) = expenditure on food by D household (rupees)
Exp. F (ND) = expenditure on food by ND household (rupees)
Cal. Intake = calorie intake per adult unit
Exp. Stat. F. = expenditure on status food, i.e. all categories of status food together (rupees)
Exp. H. Stat. F. = expenditure on high-preference food (rupees)
Exp. M. Stat. F. = expenditure on medium-preference food Rupees)
Exp. L. Stat. F. = expenditure on low-preference food (rupees)
Exp. Bat. F = expenditure on basic food (rupees)
Nut. W. = nutrition worth (rupees)
Exp. Pref. (Stat.) = Exp. Stat. F.-Nut W. = preference component (rupees)
Exp. Pref. (H. Stat.) = preference component in Exp. H. Stat. F.
Exp. Pref. (M. Stat.) = preference component in Exp. M. Stat. F.
Exp. Pref. ( L. Stat.) = preference component in Exp. L. Stat. F.

Macro-variables (Village-level)

V₁ (Eco. 1) = production profile
V₂ (Eco 2) = housing conditions
V₃ (En 1.1) = education facilities location
V₄ (En. 1.2) = health facilities location
V₅ (En. Z.1) = transport, communication information
V₆ (En 2.2) - market exposure, i.e. retail outlets for consumer goods (level and location)
V₇ (Education) ~ literacy level
V₈ (Social, cultural) . level of social and cultural events
V₉ (Size) · population of village (no.)

(All variables for years 1974, in terms of score or index. Methodology given in Appendix V₉ for 1971.)

Micro-variables (Household-level)

H. H. Size = household size (no. of members)
Dur. Assets = durable assets (score)
Curr. Borr. = current year borrowings (rupees)
Literacy = literacy (percentage of literates to total members)
F. Habit = food habit (score)
NW/W = ratio of non-working to working members
Home suppl. = supply of food from home production
Child ratio = ratio of children (percentage to total members)
Dummy (L) = dummy for agricultural labourers' households
Dummy (NF) = dummy for non-farm households
Dummy (ND) = dummy for ND

NOTES

1. In a preference space, the location of the preference curve (the indifference curve) could differ for different consumers with a given income level. When they face the same relative prices, they would demand or consume varying levels of different goods. We are not making any specific assumptions regarding the shapes of the indifference curves of different individuals other than the minimum assumptions required for the rational decision-making implied in demand theory: the curve is convex from below, and the marginal rate of substitution is negative; so that even with varying locations of the preference curve, the behaviour of consumers taken individually or together would conform to the accepted demand theory. The location of the curves could be such that consumers bunch themselves into identifiable groups other than one based on income level (a part of accepted demand theory).
2. Group behaviour, conformism, taste or preference, leadership, conventions, and even such factors as religions, beliefs, and conventions lead to identifiable groups of consumers as well as food items. Besides individual behaviour, the concept of status food or preferred food implies group behaviour leading to group preferences. In this article, however, we have extended our analysis only to identify status food and groups of consumers classified as calorie deficient and calorie-non-deficient.
3. The original data are in quantities. Common market prices are used to avoid price effects. We have little reason to believe that prices would differ greatly in the same village. If calorie-deficient families were only buyers of food and calorie-non-deficient families were producers, the price difference could be a major factor. We have shown that income and occupation classes cut across the dividing fines of the two classifications.
4. The meaning of the phrase "shadow prices of nutrients" can be illustrated briefly as follows. The required levels of nutrients, such as calories and protein, can be obtained from many items. With given market prices, a rupee (Indian currency unit) can buy certain quantities of different foods, e.g. rice and wheat, and from these quantities we can calculate the calories and protein levels. If the protein and calorie requirements are known, we shall find that a rupee can buy the necessary quantities (or more) of these nutrients from some items, but not from others, We can combine rice and wheat in various proportions to obtain the required level of calories or combine them differently to get the required level of protein. If we know the market prices of both nutrients, we can combine quantities of rice and wheat in such proportions as to obtain the required levels (or more). The cost we incur will be lower than that of any other combination of rice and wheat that could meet calorie or protein requirement. When we deal with two commodities and two nutrients the problem becomes complex, and it can be solved mathematically only through the method commonly known as linear programming (the Simplex method). The solution will pick up a basket of food that can supply the required levels of all nutrients (and more of some) at the lowest possible price; we called the commodities so selected "basic foods." From the cost of the basket we obtained the quantities of nutrients and their costs, and from these costs we calculated the shadow prices or nutritional worth of different items of food, taking into account the levels of nutrients supplied by them. These shadow prices would be less than their respective market prices since these commodities are not cost-efficient; per rupee value they supply fewer nutrients than the levels of nutrients supplied by the basic, cost-efficient items of food.
5. Both calorie requirements and requirements for other nutrients vary for different individuals. For males the calorie requirement is considered to be higher than for females; the calorie requirements of children are lower than those of adults; among adults, those not doing heavy work or those who are aged need fewer calories. Calorie requirements also vary according to body weight, climate, and other factors. In this study we dealt with families or households, each of which had varying numbers of males, females, children, and old people. Their activities also varied. Since we did not have consumption data for individual members of our household sample, we converted the members of households into adult units Consumption units) considering their sexes and ages as shown under the definition of standard adult unit (table 1). Thus we eliminated inter-household differences in sex and age. We used a requirement of 2,360 kcal, which amounts to 1.4 BIER, or 91 percent of the 1.54 BMR calculated as the requirement by the FAO for a reference man. This adjustment is arbitrary, but we have made it with the following considerations in mind. The National Institute of Nutrition considered the requirement to be 2,800 kcal for medium activity and 2,400 kcal per day for light activity for 55 kg weight. The activity differential is thus 17 percent upward (bad: fight activity). A level of 2,460 keel is half-way between 2,183 kcal and 2,800 kcal.
6. It is possible that current income for a year may differ from a permanent income level, and the latter may accord well with the economic ranking of occupations. However, we had data only for current incomes.
7. Rainfall and crop production were below normal in 1974-1975. A high percentage of calorie-deficient households at this time in Matar Taluka, a relatively developed region, could also be due to a low level of food production and possibly to low incomes during the year. Some households in the group of large farmers were badly affected by weather and suffered a 1055 of income. They would have failed to maintain their food intake level above their nutrition needs.

BIBLIOGRAPHY

Chernichovsky, D. V., and C. A. Mecook. "Patterns of Food Consumption and Nutrition in Indonesia: An Analysis of the National Socio-economic Survey, 1978. " World Bank Staff Paper no. 670. World Bank, Washington, D.C., 1984.

Food and Agriculture Organization. Fourth World Food Survey. Food and Nutrition Survey Series, no. 10. Rome, FAO, 1977.

Gopalan, C., B. V. Ramashastri, and S. C. Balasubramanian. Nutritive Value of Indian Foods. National Institute of Nutrition, Indians." ICIER Report Series, no. 60. National Institute of

Gopalan, C., and B. S. Narasing Rao. "Dietary Allowances for Indians, ICMR Report Series, no. 60. National Institute of Nutrition, Indian Council of Medical Research, Hyderabad, India, 1971.

Horton, S. "Labour Use, Nutrition and Household Behaviour: Results from Western India." Ph.D. thesis. Department of Economics, Harvard University, Cambridge, Mass., 1982.

Kumar, S. Impact of Subsidized Rice on Food Consumption and Nutrition in Kerala. Research Report, no. 5 International Food Policy Research Institute, Washington, D.C., 1979.

Marshall, A. Principles of Economics, An Introductory Volume. 8th ed Macmillan, New York, 1948.

Murty, G. V. S. N., and N. Shah. "Poverty, Inequality and Levels of Living in Gujarat." In D. T. Lakdawala, ed., Gujarat Economy: Problems and Prospects. Sardar Patel Institute of Economic and Social Research, Ahmedabad, India, 1982.

Muthaiah, C. Consumption of Cereals and Substitution of Inferior Cereals by Superior Cereals Agro-economic Research Centre for Madhya Pradesh, India, 1964.

Panikar, P. G. K. "Economics of Nutrition." Economic and Political Weekly, 7:416-430 (1972),

Pinstrup-Andersen, P., A. Berg, and M. Forman, eds. International Agricultural Research and Nutrition. International Food Policy Research Institute, Washington, UN Administrative Committee on Co-ordination, Sub-committee on Nutrition, Rome, 1984.

Radhakrishna, R., and N. Shah. "Calorie-demand Function, Price Indices, and Some Distribution Implications." Anvesak, XI (1-2): 177-202 (1981).

Reutlinger, S., and M. Selowsky. "Malnutrition and Poverty: Magnitude and Policy Options." World Bank Staff Occasional Paper, no. 23. Johns Hopkins University Press, Baltimore, Md., 1976.

Ryan, J. G., P. D. Bidinger, N. Prahlad Rao, and P. Puspamma. "The Determinants of Individual Diets and Nutritional Status in Six Villages of Southern India." Research Bulletin, no. 7, International Crops Research Institute for the Semi-arid Tropics, ICRISAT, Patancheru, Andhra Pradesh, India, 1984.

Shah, C. H. "Food Preferences and Nutrition: A Perspective on Poverty in Less Developed Countries." Indian Journal of Agricultural-Economics, XXXV (1): 1-39 (1980).

— "High Cost Calories: Food Preferences and Poverty. " Transaction/Society, 17(6): 58-61 (1980).

Shah, C. H., S. D. Sawanti, and B. T. Sanghavi. Nutrition Gap: An Economic Analysis. Himalayan Publishing House, Bombay, India, 1983.

Shah, C. H., and N. C. Shah. "In Search of Demand for Nutrition." Anvesak, X(2) (1980).

Smith, P. E "Linear Programming Models for the Determination of Palatable Human Diets." Journal of Farm Economics, 41 (2): 272-283 (1959).

Sukhatme, P. V. "Measuring Incidence of Under-nutrition: A Comment." Economic and Political Weekly, 16(23): 1034- 1036 (1981).

Theil, H. "Quantities, Prices and Budget Enquiries." Review of Economic Studies, 19(3): 129-214 (1952).

Timmer, P. C. "The Impact of Indonesian Price Policy on the Distribution of Protein-calorie-intake by Income Class and Commodity." Mimeo. Ford Foundation, Jakarta,

Timmer, P. C., W. T. Falcon and S. R. Pearson. Food Policy Analysis. Johns Hopkins University Press, Baltimore, Md., 1983.

Contents - Previous - Next