Working Paper
Adapting to Undernourishment

In estimating the prevalence of undernourishment in a region or country it has been common practice to choose a benchmark - or, as some would say, a critical limit - which reflects nutrition requirements and then to calculate the percentage of the population falling below the benchmark. (See e.g. Dandekar and Rath, 1971; Reutlinger and Selowsky, 1976; and FAO, 1978.) The logic underlying the choice of the benchmark has varied across studies. But a common driving hypothesis has been that a person's long run nutrition requirements are more or less fixed and that the variation in requirements often observed across otherwise similar people is to be explained largely by differences in their innate physiological characteristics. In other words, it has been assumed that interpersonal variations in nutrition requirements dwarf intrapersonal variations. This common hypothesis has been given its sharpest articulation in a study by Reutlinger and Alderman (1980) who, in estimating the extent of world-wide undernourishment, have dispensed with the exclusive use of overall regional benchmarks and have worked directly with a statistical distribution of individual intakes and requirements. This common underlying hypothesis has recently come under sharp attack, (Sukhatme and Margen, 1978, 1982; Sukhatme, 1981. The attack is easy to describe, but its validity is far from simple to assess. It consists of the claim that the nutrition requirement of any given individual varies in the long run over a wide range, and that this variation is achieved through an auto-regulatory process of adjustment of body metabolism. In other words, variations in nutrition intake within this range do not involve any significant alteration in the persons' weight, or in his body composition, or indeed in his physical and mental capabilities. Or, to put it in yet another way, the claim is that within a wide range a reduced nutrient intake triggers an auto-regulatory mechanism Which permits the individual to adjust - to adapt -to the 3 reduction in a costless manner. The claim is, then, that observed variations in nutrition intakes among otherwise similar persons are not to be explained by interpersonal differences in requirements, but rather by intrapersonal variations occasioned by the autoregulatory mechanism. It follows from this that existing nutrient benchmarks, or norms, such as population average requirements, as used for example by Dandekar and Rath [1971] and Reutlinger and Selowsky [197S], and person-specific requirements, as in Reutlinger and Alderman [1980], overstate greatly the extent of undernourishment and must therefore be reduced so as to encompass the fact of autoregulation.1 This claim, and the question of its incorporation in the measurement of the extent of undernourishment has caused such a furious debate in the development literature that the disentangling of rational arguments from polemics and submerged value judgements is a difficult task. Nevertheless, our concern in this article 3s to try and assess the claim. The measurement of poverty and undernourishment