Contents - Previous - Next
This is the old United Nations University website. Visit the new site at http://unu.edu
Beatrice Lorge Rogers
Supplementary feeding programmes are intended to provide an increment to the diet of a target individual within a household. usually an infant, pre-school-aged child, or pregnant woman. The food supplement also represents a transfer of real income to the recipient household. If we can assume that the household is the level at which decisions are made about the allocation of its members' resources, then the effect of the supplemental food on household as well as individual consumption must be measured in order to evaluate the programme's nutritional effect. The reason is that only by measuring household consumption is it possible to determine whether substitution is taking place between the supplemental food provided by the programme and the food that the household already consumes. If the household receiving a food supplement reduces its own food consumption by an equal amount, and uses the money freed up to purchase non-food goods, or to purchase food that costs more but has no greater nutritional value, then the nutritional benefit of the feeding programme is lost. Alternatively, the income value of the supplement may cause households to improve their overall level of food consumption, even if, as a result of leakages within the household, the programme causes no measurable nutritional improvement in the target child. Supplementary feeding programmes, then, can be of nutritional benefit in two ways: directly, by consumption of the food supplement; and indirectly, through the effect of increased income. The value of micro-economic analysis is that it can capture the indirect benefit of the programme (the income effect), while nutritional assessment of the target child captures only the direct effect.
Measurement of the income effect of a food supplement on household consumption will indicate whether the total amount of food entering the household increases as a result of the programme. It will not reveal whether the additional food (if any) is going to the nutritionally needy members; this would require assessment of the individual dietary intakes of household members. In the context of a feeding programme evaluation, therefore, economic analysis may be of limited use. Its application is limited primarily to the diagnosis of nutrition problems and their causes, and the planning of suitable programmes. If the choice of programme has already been determined, economic analysis at the household level will be of genuine use only in cases where the supplement is large enough to have a significant effect on household income.
Micro-economics is concerned with household-level consumption in relation to income and prices. Supplemental feeding programmes do not affect the food costs facing the household. (The supplemental food is free, but ordinarily the magnitude of such programmes is not sufficient to alter market prices by affecting aggregate supply or demand.) They do affect household real income in proportion to the value of the food received. Given the inevitable inaccuracy in measurements of household income and expenditure, the supplement's value, as delivered, must be relatively large to have a measurable effect on household spending patterns.
For purposes of a feeding programme evaluation, one would measure the income effect of a food supplement by comparing the total consumption of households before and after introduction of the programme, or by measuring the consumption of participant households with a sample of comparable households not in the programme. The latter approach is simpler to implement, because it requires a survey at only one point in time. It also avoids the issue of whether other circumstances changing over time were responsible for any observed differences.
If the evaluation is concerned only with measuring the programme's effects on food consumption, then only food consumption need be measured. If total household consumption is measured, then the evaluation may discover whether other forms of consumption are affected by the programme, in the event that food consumption is not.
Of course, the income effect of a given food supplement on consumption will vary by income level of the household. The measurement of income poses specific difficulties that vary according to the form and the regularity of the income received. Total household consumption is a good proxy for income and can also be used to determine what proportion of household income is represented by the food supplement.
The measurement of household consumption requires a survey approach that is relatively costly in terms of time, human resources, and money when compared with a more direct approach such as using anthropometric measurements in nutritional status surveys. If dietary recall interviews of household members are planned as part of an evaluation, the additional effort required to collect household-level consumption data may be small. In such cases, the effort may well be worth making because of the additional information on programme effects that will be obtained. If the participant population is confined to one area, and is relatively homogeneous in income and consumption patterns, then sample sizes need not be very large - perhaps 25 to 50 households in each comparison group.
Analysis of these data need not be highly complex. With a reasonable sample size, one can perform a regression analysis using the amount of food supplement received as one of several variables explaining changes in the household's food (or other) consumption. Income of the household should be one of the explanatory variables used. Proxies for income, for example ownership of key goods, may be used if income and consumption cannot be measured independently (e.g.. if most of the household's income is in the form of goods produced and consumed at home).
If a survey approach is not feasible. some effort should still be made to consider the possible income effect as part of the evaluation of a feeding programme. There are several ways this can be done. At the simplest level, it should be possible to see whether the supplemental food appears in the market place. One can probably not ask programme participants directly whether they sell the food, as they might deny it if they knew they were not supposed to do so. But one can ask whether they know how much it is worth or what they could get for it. This would indicate whether there is a market for the food and how much income it represents.
Many countries collect data on household income and expenditures for purposes of planning or for the calculation of consumer price indices. If such information is available disaggregated by income level (and by region if there are sharp regional differences in consumption patterns), it will be possible to calculate the income elasticity of demand for food for the appropriate population groups. If the approximate income level of the participant households is known, one can use the elasticity figure and the monetary value of the supplement to impute the effect of the real income value of the supplement on household consumption of food or of nutrients.
The problem with this formulation is that it is based on the assumption that an income increment will alter household behaviour in the same way, regardless of the form in which it is received. There is increasing evidence that the form of income does affect the way in which it is used (1), possibly because the form is associated with particular kinds of consumption. At the extreme, all forms of income may be interchangeable, but due to transaction costs, and to other factors that are only beginning to be investigated, they are not always treated that way. Further, supplementary feeding programmes are often associated with maternal and child health programmes that include a nutrition education component. Thus, consumption patterns may be changing separately at the same time the supplement is provided.
Another concern is that there are problems in using an elasticity calculated from cross-sectional survey data to impute consumption behaviour changes over time. To do so assumes that a low-income household, if its income rises, adopts the tastes and preferences of the higher income group it has joined. This may or may not be true. In the case of a feeding programme, though, the size of the increment represented by the supplement will probably not be so great as to put the household in a completely different income class, so the use of the elasticity is likely to be appropriate.
While these techniques are not perfectly accurate, they do provide a means for addressing the question of whether or not the feeding programme has an effect on household consumption separate from its effect on the food intake of the target child.
To analyse the income effect of a food supplement on consumption, one needs to measure income, consumption, and the value of the supplementary food received by the household. Defining the household unit is the first step in developing a survey instrument to perform these measurements. In a consumption survey. the household is conceived as the group that shares its resources for the maintenance of its members. Ordinarily, a "household" will include those living under one roof and eating together from a common kitchen. There are many cultural settings in which no group satisfactorily meets these conditions. This paper will not discuss all of the issues involved in defining the household, except to emphasize that the definition selected must be based both on a thorough knowledge of the local circumstances and on the objectives of the study - in this case, an assessment of who is likely to be affected by the income change represented by the food supplement. It should be recognized that in many cases, there will be no single respondent who can answer income and consumption questions for the whole household.
Measuring Income and Consumption
The problems of measurement are similar for both income and consumption. Both depend on respondents' recall and accurate reporting of income received or of goods purchased or otherwise obtained. Both are subject to the under-reporting bias of respondents who fear increased taxation, and the over-reporting bias of respondents who want to impress the interviewer.
Aside from any deliberate misstatement, both income and consumption are subject to the problem of faulty memory. Case transactions are easier for respondents to report accurately. In-kind payments and consumption of home-produced goods raise the double problem of estimating quantities and then of assigning monetary values. Of course, the more regular and predictable both income and consumption are, the easier they are for respondents to remember.
There are various techniques for improving the accuracy of respondents' recall. In literate populations, respondents may be asked to keep written records or notes for a period of time. Interviewers should be trained to probe for information, asking about each separate source of income or each category of expenditure or consumption. It should be recognized, though, that some errors, in measurement are unavoidable, and therefore very small differences in consumption between participant and non-participant households will be difficult to measure. Fortunately, experience has shown that with a welldesigned survey instrument and skilled interviewers, data of sufficient accuracy can be gathered to explain a great deal about household consumption behaviour.
In this context, it should be mentioned that the most accurate way to collect income and expenditure or consumption information is to question separately each earner or decision-maker in the household, and to probe for income by source. If this information can be preserved in the data analysis, it will be possible to compare the effect of the food supplement with that of other kinds of income, or with that of income flowing to different household members. This information is relevant to understanding why the supplement did or did not have an income effect, and whether alternative programme designs might have different outcomes.
The selection of an appropriate reference period is one of the issues in measurements of income and consumption. Because of seasonal variations in income, particularly from agriculture, it is usually measured on an annualized basis, though the questions should refer to the time period in which the income is received: by season for agricultural income, by week or month for wages. Consumption is subject to the same fluctuations during the year, though somewhat less so.
The reason for measuring income in a supplementary feeding programme evaluation is to determine how significant is the contribution to income made by the supplement. Seasonal variations in both income and consumption may be quite important in this context, because the food may be a significant contributor in the slack season and less so in the season of full employment, or of harvest, for example. For the purpose of programme evaluation, it must be considered whether these variations are important enough to capture in the data collection effort. In a comparison of participants with a control group, where seasonal variations are the same for both groups, a shorter reference period may be used. The season selected will, of course. affect the outcome of the evaluation.
Evaluation of In-kind Transactions
The evaluation of in-kind transactions in monetary terms poses a problem for which there is no perfectly satisfactory solution. If the in-kind goods are also marketed, the value is equal to the market price. In the case of goods produced by the household, the wholesale price should be used, as this represents the income that the producer could have received for the goods. But as prices vary seasonally and regionally, one must select the price at the appropriate time of year and in the market closest to the area being studied. For goods received as wages, the retail price is more appropriate, because it represents what the recipient would have had to spend for them. While some discount should in theory be made to account for the recipient's lack of choice, the full price is usually used. Some goods are not traded. Some kinds of food that are foraged, for example, never appear in markets. In such cases, the usual solution is to use the value of the closest substitute that does appear. This selection is a matter of judgment and will not be a perfectly accurate reflection of the true value of the good.
Evaluation of the Food Supplement
Determining the value of the food provided in a supplementary feeding programme is simply a special case of evaluating in-kind transactions, but it poses a few special difficulties. Food supplements are often specially formulated foods that do not have an exact equivalent in the local diet. They are often fortified so that their nutritional value exceeds that of the closest local substitute food (which may indeed be the local staple grain). Particularly if a programme evaluation is performed by the agency running the programme, there may be an inclination to overvalue the food to account for this high nutritional content. This is inappropriate to an analysis of the income effect, unless the local market also takes account of the increased nutrient content of the food, because it is the monetary value to the recipient, not to the provider, that will determine the income effect.
The usefulness of micro-economic analysis in the evaluation of supplementary feeding programmes depends on the purposes of the evaluation and the resources available for it. If the evaluation is concerned only with quantifying the programme's effect on the nutritional status of the target child, then economic analysis need not be performed. Economic analysis is necessary, however, to determine the income effect of the food supplement on the consumption of other household members, which may be an additional nutritional benefit of the programme. Further, it is critical to understanding the mechanisms by which a feeding programme achieves its effect, so as to predict the effects of programme modifications and to compare feeding programmes with alternative uses of available resources.
Measurement of income and consumption is subject to errors of deliberate misreporting and of faulty recall. Some error is controllable by means of careful design and implementation of the survey, but some degree of error is unavoidable. The income effect of a food supplement must therefore be fairly significant if it is to be measured in an evaluation. The evaluator must make a judgment on this based on the kind and quantity of food delivered to recipient households and their income level.
The survey approach used to obtain household economic data requires relatively trained manpower and is more costly and time consuming than other evaluation approaches, though it will be much less expensive if it is an add-on to a household dietary survey. Even a small sample of households can provide valuable information on the response of household consumption behaviour to receipt of a food supplement. If such a small-scale household survey is not feasible, this chapter has suggested that observation of local markets and possibly analysis of existing information may still provide insight into this behaviour.
Whether or not a programme evaluation includes primary collection of micro-economic data, an understanding of the possible economic effects of the programme will result in better informed and more reliable conclusions.
Contents - Previous - Next