Cover Image
close this bookThe Himalayan Dilemma: Reconciling Development and Conservation (UNU, 1989, 295 pages)
close this folder7. The human dimension: what are the facts?
View the documentIntroduction
View the documentPoverty in the Himalaya
View the documentWomen and young people
View the documentConclusions

Poverty in the Himalaya

This section draws heavily on one of the Mohonk Mountain Conference background papers which discusses poverty, women, and young people in the Himalaya (Pitt, 1986). It analyses information available in the files of major agencies of the United Nations Organization and the World Bank in an attempt to lay a foundation for answering the question: what is poverty in the Himalayan context and how is it determined?

Any large agency that has to face up to responsibility for a major section of humankind, will have to address the problem of data collection, storage, retrieval, and analysis. This is especially formidable when dealing from scratch with such issues as health, poverty, human welfare, in Developing Countries that were initially without any centralized bureaucracy comparable to those that inhabitants of the industrialized countries accept as given. Thus problems of data reliability, even of what kinds of data to collect, are a part of the process from the beginning. It is appropriate, therefore, that we attempt a review of this process in the context of the Theory of Himalayan Environmental Degradation.

For some decades the World Bank has ranked the countries of the Third World using a variety of socio-economic indices. From this we are informed that, amongst the large number of very poor countries (Least Developed Countries = LDCs), Nepal and Bhutan were ranked fifth-to-last and last, respectively, in the 1981 World Bank Atlas. We cannot assess the comparable standing of other sections of the Himalaya (for example, the Kumaun and Garhwal Himalaya of Uttar Pradesh State, India, relevant parts of Xizang Autonomous Region, People's Republic of China, and northern Pakistan) because the relevant data are aggregated within those of the respective nation state. We will assume that, if specific data had been available, these mountain regions would compare closely with Nepal and Bhutan.

The ranking of Nepal and Bhutan is based upon a comparison of the Gross National Product (GNP) per annum per capita. The five bottom states on the World Bank LDC list are regarded as being in a condition termed 'supercrisis' (we intend to use the term'supercrisist on a regional scale here, rather than at the level of a single country, and will discuss it in more detail in Chapter 9). Since the other three lowest rank positions (in addition to Nepal and Bhutan) were taken in 1981 by Bangladesh, Chad, and Ethiopia, this alone should indicate the seriousness of the Himalayan situation if we assume that the listing has any significance.

Relying for the moment solely on the inferences that have been drawn from ranking by GNP per capita, it appears that extreme poverty characterizes subSaharan Africa and the land-locked areas of the Himalayan region, plus Bangladesh, and several Andean and other countries. Furthermore, the most recent figures from the World Development Report (World Bank, 1984a:218) show little relative change from the 1981 listing. There is perhaps one significant exception: Bhutan has dropped off the list entirely. Nepal, after showing an increase in the growth rate of GNP per annum per capita in the 1981 Atlas, registered in 1986 an annual decline of 0.1 percent for the period 19601982 - a rate of decline only exceeded by Chad amongst the last five LDCs.

Pitt (1986) makes the mildly cynical suggestion that most of the new Himalayan data on poverty refer to Nepal since it appears to have become very fashionable to study Nepal, both for academics and for agencies. This, if justified, raises another issue relating to data reliability, comparability, and representativeness, and is worthy of a fuller discussion. The reasons for going to Nepal, of course, may in part be due to the beauty, romance, and general attractions of the country and the friendliness and hospitality of its people. Pitt insinuates, however, that it may be due also, at least in the case of some agencies, to the fact that opportunities for intervention elsewhere have been significantly reduced. The favourite countries for technical aid in the 1960s have since turned sour. For instance, at one point the United Nations was even evicted from that South Seas Paradise, Western Samoa, while in many African countries military coups and dictatorships have made life very uncomfortable for development and research agencies: on one occasion a top-level UN mission on rural development was detained in a hotel for several weeks. Within the Himalayan region itself university scholars certainly find access into Nepal much easier than into most other areas: the northern frontier zones of India are especially restricted.

This openness of Nepal, however, may be as illusory as the welcome to those outsiders who thought that they had grasped the Samoan situation. The waters of Nepal may be even more muddy and the cultural mosaic more intricate. Certainly, the presence in Kathmandu of a large expatriate subculture living on comparatively enormous salaries and the lack of coordination amongst a bewildering array of UN and bi-lateral agencies and competing projects, have caused a development problem of their own. And as we implied earlier, with certain exceptions, much of the rest of the Himalaya are virtually a closed book, and even the new UN reports on Bhutan are not very informative.

The ranking of LDCs, therefore, and especially the very low ranking of the Himalayan region as a whole, must be viewed critically in terms of data reliability. The indicators of poverty that are normally used, GNP per capita as indicated above, and social indicators, such as infant mortality rates, do not vary significantly over time, nor among the different sources in which they appear. But this may be because the sources are all quoting the same primary data so that the appearance of corroboration is spurious. Nevertheless, even if the data do need to be handled with scepticism, some relative points can be made.

There are two groups of comparative indicators of poverty which are in wide use, the one economic, the other social (notably health and education). In the 1984 World Development Report, two economic indicators are termed basic: the GNP per capita and the inflation rate. The Nepal score in the first is US $170 per annum per capita for 1982. This shows an increase of US $40 over the 1979 figure (World Bank Atlas, 1981). That this is regarded as a decline by the economic wizards is apparently not sleight-of-hand but the product of the method of calculation of GNP (World Bank, 1984a: 274), and the fluctuating rate of the purchasing power of the US dollar.

It is generally agreed that there are many things wrong with using GNP as an indicator of poverty, or indeed of development. This is especially true when we descend from consideration of a country as a whole to that of an ethnic minority, or of an individual family. GNP is a measure of the total domestic and foreign output claimed by residents, without any reduction for depreciation, divided simply into the total population. No distinction is made for distribution of income. It is not possible to include the so-called 'black economy' or the activities of big, and especially multi-national, companies, nor the extent of foreign-exchange transactions. Only quantified and quantifiable elements are included, so that in addition to the other limitations, the entire subsistence economy is omitted from consideration. In Bhutan, the 1981 World Bank Atlas claimed a GNP per annum per capita of US $80 for 1979. This figure is quite meaningless because it is estimated that traditional subsistence agriculture accounted for more than 90 percent of total 1980/81 production (World Bank, 1984b: 18). Mr. D. N. S. Dhakal (personal communication, 1985), a UNU Fellow working with us, made the telling point that in his village of birth in Bhutan the average farmer, who is the faceless statistic accounting for the average US $80 GNP per capita/yr, would not know what to do with the money if he had it -'it is GNH, gross national happiness, that counts, not GNP!' A similar situation prevails in Nepal and presumably throughout the Himalayan region. Gross National Product per capita, in terms of access to cash, hardly indicates wealth or poverty in a nonmonetary economy judged on local standards rather than from a Eurocentric perspective.

An interesting issue is the starting point for the entire process since the GNP per capita for Year I, the first year that a country is ranked, provides a benchmark for evaluating its economic progress or decline. The case of Nepal is of particular interest and it may serve as a caution when viewing the ranking progress of several other countries. In the early 1950s Nepal was virtually bereft of normal government files and accumulated information relating to the national economy. Stiller and Yadav (1979: 35-36) make the point that: 'It is common knowledge, for instance, that the base-line for economic growth, the per capita income for Nepal during the early years of the planning period was arrived at by just such a method of approximation.' They go on to recount Dr. Y. P. Pant's account of this amazing story. It is quoted in full, not only for its own sake, but because it also explains how he, as Secretary of Finance, Secretary of Planning, and Governor of the Rastra Bank, chose to qualify the deed.

In Nepal no serious attempt could be made earlier to calculate national income due to the dearth of statistics. The first attempt was made by Prof. W. O. Thweatt, an economist in the Ford Foundation, for the year 1954. For 1954 he estimated the national income . . . at $40 (Rs 304) [per capita]. He derived these figures simply by comparing the per capita income of Pakistan, which he guessed to be almost similar to Nepal's with respect to economic development during the period.... Prof. Thweatt's estimate was almost an ad hoc approach, which, no doubt, has helped to create some basis for further investigation. He himself calls it only a rough approximation. (Y. P. Pant (1970) quoted in Stiller and Yadav, 1979:36)

The commentary by Stiller and Yadav that follows this quotation is also worth citing in full:

Besides explaining the origin of this statistic that is often quoted for establishing growth curves in Nepal, Dr. Pant has provided some interesting examples of how even among economists statistics slip from 'rough approximations' to harder facts. He calls Thweatt's estimate almost an ad hoc approach, even though he knew very well that it was sheer fabrication. (Stiller and Yadav, 1979:36)

Stiller and Yadav also explain that Thweatt's knowledge of Nepal was 'good for a man in his position but limited nonetheless.' He had absolutely no observational data and so was prompted to make 'this hilarious jump to Pakistan ...,'and Dr. Pant, after sombre reflection on this 'almost ad hoc approach,' believed that it assisted in creating some basis for further investigation. The most telling point, however, in this passage from Stiller and Yadav (1979) is their concluding remark: 'Even the softest of statistics are treated progressively as harder and harder data until the statistics become accepted as fact.' This at least draws the useful comparison with our earlier concern over the physical 'facts': the arena we have now entered should not be entirely unfamiliar even though we are facing a different species of lion.

To conclude the excursion into establishment of Nepal's benchmark GNP, if we accept Dr. Pant's approach, then we have demonstrated that by 1982 Nepal's GNP had improved by a factor of more than four. If we refuse to accept the approach, what else is available? This is the very problem that faced us in the preparation of this chapter: if we leave our readers confused over which conclusions or inferences to accept and which to reject, at least they will sense the essence of the dilemma that faces all decision makers and most scholars.

To be fair, the World Bank has drawn attention to the fact that GNP per capita figures should be used with extreme caution. And in general they are supplemented by a battery of other economic and social indices. For example, the 1981 Conference on Least Developed Countries (UNCTAD, 1981) omitted GNP per capita completely and placed more weight on agricultural indicators (such as percentage of labour force in production, agricultural production growth rates, output per worker, and so on). The agricultural production growth rates for Nepal showed an actual I percent per capita decline for the period 1960-79. While the Bhutan growth rate was zero, this is better than the average rate for all LDCs which showed a decline of 0.8 percent.

One indicator, however, is quite remarkable for Nepal: that is the 1976-77 income distribution figure quoted in the World Development Report for 1984 (World Bank, 1984a: 272). According to this, the highest 10 percent of households accounted for 46.5 percent of the total income. This is very nearly a world record, being exceeded in the World Bank table only by Brazil with 50.6 percent. This means that the distribution of wealth in Nepal is highly stratified and that poverty is heavily concentrated in the lower-income groups, creating a large gap between them and the upper-income groups. Of course, it is another reflection of the predominantly subsistence economy; and it would be instructive if comparable data could be obtained for various Indian states which have a lowland and a Himalayan division, such as Uttar Pradesh, Himachal Pradesh, Jammu and Kashmir.

Inflation is a much more difficult indicator to use in the Himalayan situation although there are certainly figures available. For example, the World Bank 1984 [Report for Nepal gives a rate of 8.9 percent per annum for the period 1970-82, compared with 7.7 percent for 1960-79, and 11.7 percent for the thirty-two low-income countries (excluding India and China). Nevertheless, the rate for Nepal had exceeded 10 percent by the end of 1983 (ESCAP, 1984). The 1984 World Bank Bhutan Country Report suggests a rate of about 11 percent for consumer prices. The figures for Nepal and Bhutan are relatively low compared with some African countries where inflation is running away at 30 percent or more per annum.

We must ask, however, what do inflation figures indicate in terms of purchasing power or, more precisely, the ability of an ordinary family to feed itself and provide for basic needs? The lowest average inflation reported by the World Bank for any country in the world for the period 1970-82 was Ethiopia! By the end of this period, as is well known, Ethiopia was facing a national calamity and earned the descriptor of 'supercrisis' state.

The inflation rates quoted refer only to quantified elements in the commercial economy and tell us little about the subsistence economy. In the case of Bhutan, with over 90 percent of the economy estimated as subsistence, the inflation rate is merely a mirror image of the cost of imports from India. The subsistence sector in Nepal is probably between 70 and 85 percent, although it is declining (Acharya and Bennett, 1983, Table 2).

Inflation, of course, can take place under subsistence conditions, as anthropologists have shown in discussing the Kwakiutl potlatch, a traditional feast, and similar conditions. More relevant to the Himalayan region, however, is the high and rapidly increasing cost of marriage (Jiha) in Hindu society, and of the arghun funerary activities and expenses amongst the Gurungs (Messerschmidt, 1976a). But in these situations there are ritualized means of stopping an upward spiral of inflation and of mitigating any effects, notably through gift exchanges. The basic problem facing subsistence farmers, however, is that people cannot transfer their subsistence incomes (farm and forest products) into cash, nor can they easily obtain cash. This becomes more serious when taxes and other financial obligations must be paid in cash rather than in kind. Therefore, the inflation rate is effectively increasing the difficulties of obtaining any cash, and thus enhancing the processes which divorce them from their land, livelihood, and means of self-reliance. If it could be calculated, this inflation rate likely would be very high indeed.

The difficulties in drawing meaningful conclusions from income figures and other economic indicators have led to an emphasis on social indices - the socalled physical quality of life indices (PQLI) - a step closer to the Bhutanese gross national happiness evaluation. For example, some countries which rank very low in the World Bank's (1984) GNP per capita table (Western Samoa, Sri Lanka, the state of Kerala, India) show a much higher standing when the indicators used are infant mortality rates, literacy (notably female literacy), and life expectancy. Health without wealth has been seen as a desirable goal for development and a situation where poverty (as defined by the numbers game being discussed) is not necessarily a problem at all. Such a circumstance has been variously explained (Defence for Children, 1982). The status of women and a relatively egalitarian social structure seem important prerequisites. For example, in income distribution tables, Sri Lanka scores very well (28.2 percent of the household income being obtained by the highest 10 percent of households, an egalitarian rate compared to some of the other LDCs).

The PQLI figures, however, are also suspect, especially infant mortality rates, since many people do not register births and certainly not the infant deaths, even discussion of which is often taboo. Without these figures, expectation of life can only be a guess. Literacy, also, is an ethno-centric concept, and is based upon criteria obtained by written cultures (usually the industrialized countries). Most people, especially those of the LDCs, are part of the world of oral cultures.

With all of these reservations, let us consider the PQLI for Nepal and Bhutan. They still do not indicate an acceptable standard of living, although they are not apparently declining. For example, infant mortality rates in Nepal for 1980 were given as 150 per thousand (UNCTAD, 1981:32ff) whereas the World Health Organisation (WHO) goal, for comparison, is 50, and the average for all low-income countries is 87 per thousand. The 1982 figure for Nepal was 145 per thousand (World Bank, 1984a:262). For Bhutan it was given as 147 for 1982 (World Bank, 1984b:18). The adult literacy rate in Nepal in 1975 was 19 percent, although the most recent government figures (1981) claim 24 percent (Ministry of Education and Culture, 1984), which is the LDC average. The adult literacy rate in Bhutan for 1982 was set at 10 percent (World Bank, 1984b: 18).

Access to safe water has been a major problem, and probably underlies the high infant mortality rates. In Nepal in 1975 it was calculated that only 9 percent of the population had access to safe water (UNCTAD, 1981:32), and the experience of tourist visitors to Kathmandu might serve to challenge even this low figure as optimistic. In Bhutan the figure is probably lower, as only 6 percent of the rural population was rated to have access to safe water in 1982 (World Bank, 1984b:18) compared with a 35 percent average for all LDCs. In contrast, both Nepal and Bhutan had a higher than average daily calorie intake as a proportion of estimated minimum requirement (91 percent and 88 percent, in 1977, respectively), compared with 86 percent for all LDCs (UNCTAD, 1981:32). An overall assessment of these PQLI figures does indicate that there is some improvement, provided we can assume that the margins of error are consistent, and this is not necessarily a safe assumption.

The conclusion may be drawn that, by available measurement standards, Nepal and Bhutan are very poor countries though in some areas (notably water power) there is a great potential for improvement (World Bank, 1979). The assessment of great poverty, however, does not necessarily infer that they will become supercrisis states like Ethiopia and Chad. It could be argued that in those latter two countries there are additional catastrophic elements that are not present in the Himalaya et the moment. In Ethiopia disaster, including a rapid rise in mortality, is due, in part, to political unrest and conflict in the midst of drought, which not only takes its own toll of lives but also disrupts subsistence agriculture and severely curtails the effective distribution of food aid. Nor is drought a particularly pervasive and significant phenomenon in the Himalaya, as it is in Ethiopia and Chad and other parts of Africa.

In order to understand better the relationships between the environmental problems in the Himalaya, however defined, and poverty, and the effects of recent changes, it may be necessary to disaggregate, to look in greater detail at the poorest groups within the region, and at changes over time. Here the UN and World Bank statistics are not particularly helpful, partly because they consistently aggregate, and partly because quantitative figures may be lacking for certain areas and for time series, and because they tell us little about causes and dynamics. Much more reliance, therefore, has to be placed on field studies including, where available, anthropological work, even if this work is scanty, and may not be at all statistically representative.

What, then, can be said about recent changes in the poverty situation and the links to the broader environmental problems? What light is thrown on the key causal patterns? The simple explanation proposed by many outside agencies has been population pressure, and indeed indicators based on per capita estimations suggest that poverty deepens with increased population. One might presume, however, from one statistic that this is not the way women, at least, interpret the situation. The World Bank's 1984 Report (p. 256) states that in Nepal only 7 percent of married women of child-bearing age are practicing contraception (itself one of the more dubious statistics). The population appears to be increasing at an unexpectedly rapid rate. Nor is everyone blaming overly fecund peasants. One recent International Labour Organization (ILO) study has boldly stated that'the conclusion often drawn that it is population growth that is largely responsible (their emphasis) for the growing poverty of the mass of the Nepalese people - is a gross oversimplification and ignores the crucial fact that demographic change, like material deprivation or poverty, is a social product, conditioned and determined in the last analysis by the economic and social structures of the state' (Seddon, 1983: 1). Population growth, then, may be regarded as one of the symptoms, not one of the causes, itself to be explained by what is happening in other sectors.

The Seddon ILO (1983:32) study is of exceptional interest since it is based on detailed fieldwork of household behaviour and economic status. A major emphasis is food production - not the overall quantities but the access that households have to food supplies. The conclusion is that grain availability for Nepal had moved from surplus in 1971 to deficit in 1981 (from +66,921 to 108, 278 metric tonnes). In 1981 the only region with a surplus was the Terai and it was predicted that this would be eliminated by the late 1980s. The Terai surplus so far has been earning foreign currency rather than being used to maintain minimal nutritional standards. Messerschmidt (personal communication, March 1987) sees the surplus grain of the Terai as earning two things: (1) foreign currency (initially); and (2) foreign debt (incurred later when Nepal re-imports rice to meet deficits, usually in the Spring, when local stocks have been depleted). The situation in the Middle Mountains is very much worse, and the 70,697 metric tonnes deficit of 1971 increased to 123,755 metric tonnes in 1981. In the Terai both production and yield, according to the Ministry of Food, Agriculture, and Irrigation, seem to have declined between 1974 and 1978, despite the increase in land under cultivation (60,000 ha) through irrigation, the rise in fertilizer use of about 40 percent, and the enormous increase in area under high-yielding crop varieties (Seddon, 1983:34). The reasons given are poor weather, failure to extend irrigation further, and inability to apply new agricultural technology. All crops seem to be in decline, except possibly rice (Seddon, 1983).

These conclusions of Seddon, based upon a historical perspective, are now somewhat out of date as well as being open to challenge because they were derived from fieldwork in a limited area. Since the 1970s wheat production in the Middle Mountains and in the Kathmandu Valley has increased significantly, due in large measure to the introduction of new varieties and the rapid growth in cultivation of winter wheat. Potato production has also risen, although its contribution to the total food consumption is rather insignificant (Horton, 1987). To cite one specific instance (Messerschmidt, personal communication, March 1987), Marsiangdi, Lamjung district, became a winter-wheat producing centre of some note in the ten years following Messerschmidt's introduction of new varieties in 1972. Today the entire valley is green with wheat in winter when previously it was brown with fallow.

Such improvements are to be found in many parts of the Himalaya. Nevertheless, food shortage crises, especially in the Middle Mountains, occur with increasing frequency (it appears that the improvements do not keep pace with population growth). As a result, not only are foreign earnings from export crops declining, but international food aid is required. A special committee was established in 1980 and agreements were signed with the World Food Programme, India, USA, and others. Nevertheless, food has continued to be exported and new agreements have been reached with China, for instance, for the export of flour, beans, and pulses. Only some mountain regions maintain grain surplus, but even here food aid and subsidization has been necessary.

The implications of this combination of local food deficit and increased aid are considerable. It has been shown that the effects of such aid in Africa, where there have been many negative development effects, such as corruption and internal distribution problems, a dramatic decline in subsistent, self-reliant, agriculture, and a decline in maternal and child health, have occurred (Defence for Children, 1982). A similar scenario seems to be building up in Nepal.

The aid-giving agencies, however, have fought vigorously against the notion that introduced aid may be counterproductive. The crisis of food production in Nepal has been blamed on the Nepalese farmers who, according to several agencies, have not accepted quickly enough the introduced technology, but have preferred to stay with what is called their 'archaic' and 'traditional' agriculture. Such a stereotype needs careful examination. Recent studies (Seddon, 1983) have shown that the larger farmers with more land (especially in the Terai), more financial resources and capacity, do invest in the new technology. Tractorowning farms were an average size of 17.7 ha, compared to a general average 3 ha according to one study in Parsa and Chitwan. But even here, investment was modest by international standards. One can argue, however, that machines have a limited utility on slopes, and the cases cited here refer to the Terai.

Fertilizer use is limited, according to Seddon (1983), not because of lack of finance, access, nor the weak development of agricultural extension services, but by the farmers' own choice. Nevertheless, the large increases in wheat production, referred to above, result primarily from the introduction of new varieties and the adoption of winter cultivation and reduction in winter fallow, rather than from increased use of artificial fertilizers.

The lessons from the history of agricultural production are relatively simple. The retention of traditional technology has not only a long-term logic in the sense of preserving a delicately balanced agro-ecosystem, but is probably the only available course of action when incomes, notably from the agricultural sector, fall below a certain level (Pitt, 1970, 1976; McNeely and Pitt, 1985). At this point, population increases because traditional agricultural methods are labour intensive (this is only one of a number of possible causes of population growth). It is human hands and energy that provide the capital. In this respect, there appears to be a striking difference between Nepal and Bhutan, the socalled Himalaya poor states, on the one hand, and the supercrisis states of Africa, on the other. Traditional African agriculture or pastoralism have been seriously undermined and in places virtually destroyed, in some cases by deliberate government policies which have discriminated against marginal social groups, and have, in the interests of centralized political control insisted on sedentarization (see also Blaikie, 1985; Blaikie and Brookfield, 1987). The traditional Himalayan social system cannot be broken so easily in this way, although one might hypothesize that governments may have a vested interest in encouraging, or at least in not discouraging, the recent urban drift, because it provides a reserve army of labour and may facilitate political control.

It is already obvious that any simple label of poverty does not adequately define a situation where there are very considerable sociological differences in the degree of poverty. What are these sociological dimensions in Nepal and other areas of the Himalaya? There are some interesting recent materials from which a more detailed mosaic can be constructed, although these deal mainly with Nepal.

A first fact is land ownership. Generally speaking, those who own land are better off than those who do not, and the larger the landholding the higher the standard of living. The World Development Report, World Bank (1984a), as mentioned above, indicates a high degree of concentration in the distribution of Nepalese wealth. In 1981, in the rural sector, 2 percent of all rural households were said to cultivate about 27 percent of the land (Seddon, 1983:93). The landreform measures had obviously not worked, partly because the official ceiling for landholdings was quite high, 25 bighas (about 18 ha). Inequalities in size were most marked in the Terai, but in the Middle Mountains the problems continued to relate to the small, often uneconomic size of holdings. However, the proportion of landless (continuing at about 8-10 percent) is not high by Asian standards, but most of the rural poor who are landowners have too little land on which to eke out a viable living. They are therefore forced to rent additional land, to practice sharecropping on larger estates, or to seek waged labour. These options may be precluded, at least locally, by the high rentals and the scarcity of employment opportunities. As a result, according to Seddon (1983:93), 63 percent of family labour days were underemployed. This has led to a further stimulation of migration.

The size of landholdings also relates directly to nutrition and calorie intake. A 1978 survey of households showed that farms of less than I ha had an average calorie intake of 1,500, or less, per person, whilst those over 1.01 ha produced 2,250 calories per person, or more. This depends to some degree on types of crops and the annual yield but overall, according to Colin and Falk (Kosi Hills Rural Development Project - KHARDEP, 1979), small farms are deficient, and farms under I ha produce only 60 to 70 percent of household requirements while those over 1.01 ha produce 86 percent.

The Agricultural Projects Services Centre (APROSC, 1978) report that annual per capita food consumption (in kg) varies significantly between regions (297 in the far west, 480 in the centre, and an average of 373 for all Nepal). In the far western region 59.2 percent of households were considered to be below the poverty line (defined in Nepal as an earning capacity of 2 N rupees/day on 1977 figures). What is not clear from these admittedly inconsistent figures is how quickly the situation may be deteriorating. A decline is not necessarily corroborated by the steeply rising migration rate because there are many reasons, other than economic, for migration (Pitt, 1983), but for the Himalaya, the two are probably linked (Gurung, 1973; Elder, 1974; Manzardo et al., 1975).

Landholding, in certain circumstances, may be a sign of poverty. Nevertheless, out-migration from the Middle Mountains is usually the result of abandonment of marginal land which cannot provide the minimum sustenance for enlarged families. However, since there is no base-line it may be that the numbers of those below the poverty line (which is more of an ethnocentric concept, rather variously applied from outside) have not changed greatly.

What is more interesting perhaps in the most recent figures is the indication of the concentration of poverty. But it is significant that the figures do not pinpoint the social groups which are most deprived nor explain why. Individual regions contain many ethnic and caste groups, households vary in sex and age structures, and the size of landholding tends to become irrelevant when the vital extra income has to be earned by other means. Seddon (1983) did not find an answer in any feudal-type obligations since, presumably, even the larger landowners had few resources. More significant may be the differential access which different castes and ethnic groups have to wage-earning possibilities: for example, Gurungs, Magars, Limbus, and Rais are noted in military service, highcaste Brahmins and Chhetris are preferred by employers, and 'untouchables" encounter employment barriers by virtue of their low or outcaste social status. These conclusions of Seddon (1983) must also be handled with extreme caution, as the whole issue of employment opportunity is much more complex than one of caste and status. Status plays a role, along with the related factors of education and political influence. It could be argued that 'untouchables' encounter no greater employment barriers than others in the sense that (a) the upper castes have cultural barriers to many categories of employment also, while (b) lower castes have access to manual labour, portering, etc. from which the higher castes are barred. The argument could even be turned on its head: high castes have reduced latitude for employment because of the restrictions of their caste status (Messerschmidt, personal communication, March 1987). Nevertheless, there are widespread indications of a relatively steep recent fall in standards of living because of the rural people's declining purchasing power (for rice or maize, particularly) (ARTEP, 1982).

This fall in purchasing power seems to have had a number of major effects in terms of the satisfaction of basic needs. First, nutrition: in the Middle Mountains and high Himalaya total available calories per person per day dropped from 1,569 in 1977-77 to 1,426 in 1983-84, and they were projected to drop to 1,299 by 1989-90 (Seddon, 1983:124). Though the situation in the Terai was not as critical as this in terms of averages, most people had less money, especially the smallest farmers and the landless. Nutritional problems were accentuated amongst certain ethnic and caste groups because of food taboos. The fact that demand concentrates on certain foods raises the price for these foods, although caste and ethnicity may be less important in some areas. Sacherer (1979) has shown clearly in her studies in Dolakha district how this factor affects food sufficiency. Here Brahmins had from 7 to 9 months of food sufficiency whilst 'low castes' had only 4.6 to 5 months (other figures were Chhetris: 5.9 to 6.9; Newars: 6.8; Tamangs: 8.4; Sherpas: 4.8).

This example indicates how the problem of malnutrition, at least, is in part a result of a deleterious change from traditional customs where taboos were in many senses related to seasonal supply and demand, and the 'modem' system which has superseded it, where supply and demand are manipulated or otherwise work against the consumer. The pressures on traditional society, deriving in part from progressive intrusion of the world monetary economy, therefore, may have accentuated the problems.

The unwise change from traditional management also applies in time as well as space. The months prior to the monsoon are months of hunger. To bridge the food-gap, people such as the Magars, Gurungs, and Tamangs, for example, have traditionally foraged for forest products at this time: tubers, berries, and such like, and have taken game birds, animals, and fish. The deteriorating quality of the forests and the difficulties of common-land rights, have further curtailed these possibilities. People may even eat the next year's seeds at this time. More significantly, this is the time they fall into debt. The question of indebtedness, however, is also very complicated and will be discussed further in Chapter 9. Here, nevertheless, it is necessary to indicate that changing (rising) aspirations play a large role. For instance, increasing desires to own "luxury" goods such as radios, wrist-watches, and to enable at least one child to go to school, all place an increasing strain on very limited monetary means, leading to indebtedness.

There are, in fact, estimates that 90 percent of all debt in Nepal is for consumption items (Seddon, 1983:132). As much as 80 percent of land (especially irrigated land) may be pledged (Caplan, 1970). Interest rates may be extremely high. In the Middle Mountains the problem is particularly acute, with reports that 35 percent of households are in debt. On the other hand, the debts may be rolling ones, where the principal is not repaid and foreclosure rare. Debt, of course, also occurs in the traditional sector, notably for life-cycle ceremonies. Here, however, it was a redistributive mechanism, part of the Maussian gift-exchange structure of reciprocity, where giving (and lending) was a prerogative of rank. This traditional ethos may have carried over into the modern situation, so that debt is perhaps not the enormous problem it appears to be from the available figures. Rotating credit associations are other local mechanisms that have been established to ameliorate this problem (Messerschmidt, 1978).

The effects of malnutrition, however, are cumulative. For example, there is a vicious circle between decreases in production and increasing malnutrition. An outcome is ill-health, particularly the incidence of debilitating disease and, notably in Nepal, diarrhoea. Blindness may be a result of malnutrition, especially Vitamin A deficiency. On the other hand, most health indicators, from infant mortality rates on, show a gradual improvement, as the number of physicians per capita has more than doubled in the period 1960-80 (World Bank, 1984:267). But even this may be mainly an artefact of collected statistics, such as the number of hospital admissions, which refers overwhelmingly to the Kathmandu region. There is much less information for other regions, or social groups, and what is available is often contradictory.

The available evidence suggests that any discussion of poverty and basic needs in relation to wealth, income, and land, needs careful interpretation. For example, Sacherer (1979), in her studies in Nandu Kabre, has pointed out that whilst wealth may increase as one ascends the caste ladder, health, or at least nutrition, may decline. The main reason is that 'the higher the caste group, the greater the number of restrictions on diet' (Sacherer, 1979:31). These restrictions constrain what is eaten by which member of the family, what is eaten at what age, by women or men at what time of the month, year, and so on. Moreover, the more able and ambitious people, who wish to acquire wealth and status, will restrict their diet to move up the social ladder. High-caste children, Sacherer concludes, may therefore suffer the greatest malnutrition. Brahmin food prohibitions, for example, include beef, buffalo, pigs, chicken, and eggs. Children eat after adults in high-caste families, and girls after boys. If there is a shortage of food the child at the end of the line suffers most. There is a Nepalese saying that 'boys are raised up, girls are starved up' (Messerschmidt, personal communication, March 1987). It is possible then to have wealth without health as well as health without wealth. But even these social correlations should be handled with care. For example, many young Brahmins told Sacherer that they did not believe in caste rules, restrictions, or privileges. They had travelled to Kathmandu and, indeed, often to India and, whilst orthodox in their home villages, wanted changes especially for their children. The situation was complicated by inter-caste and inter-ethnic marriage (for example, a Brahmin married to two wives, one Brahmin and one Chhetri). It is uncertain, however, whether social distance affects inter-marriage; in other words, do proximate caste people inter-marry more readily, and how often is caste inter-marriage part of wider Maussian gift exchange where things cancel each other out? Nor is there an adequate explanation of the 'success' stories that Sacherer (1980) encountered among the mid-ranking Sherpas, Jirels, and Newars. Finally, Sacherer's study is only of two panchayats and it is not known how representative these are.

Some speculation is in order here. As an example, there is the curious paradox in Sacherer's figures that the groups which made the greatest use of the health posts - the Brahmins, Tamangs, and Chhetris - had the highest rate of child malnutrition, although the relationship to mortality and morbidity rates is unknown. The health posts also may not be visited by lower-caste people and ethnic minorities where they are typically staffed by high-caste people, based in part on their easier access to education. The hypothesis proposed elsewhere (Defence for Children, 1982) that health posts may do more harm than good (for example, by distributing inappropriate baby foods) should be carefully examined in Nepal. Another interesting point is that female child malnutrition rates were higher than male in all caste and ethnic groups except the Sherpas and Jirels where the rate for males was higher. This would seem to indicate that females are not prejudiced against amongst the Sherpas and Jirels, as they appear to be in many caste and ethnic groups. This raises several questions. Once again, is the traditional context important in improving health and nutrition? Is the introduced system of health, in some senses, counterproductive?

Discussion of other basic needs, such as education, housing, and water, is restricted by lack of evidence. Sacherer's (1980:18) studies indicate that Brahmins in one panchayat were considerably more literate (23 percent) than Chhetris (12 percent), or Newars (9 percent), but not Tamangs (22 percent). But in the second panchayat Brahmin literacy was both low (8 percent) and not significantly different from that of Sherpas and Jirels (7 percent). Overall in the Dolakha district males were much more literate (18.3 percent) than females (2.3 percent). The national averages at that time (presumably 1977) were 23.6 percent and 3.9 percent, though the most recent survey of education (Ministry of Education and Culture, 1984) gives males 34 percent (six years and over) and females 12 percent. Brahmins and Tamangs were also over-represented in the primary-school population of the two panchayats, whilst Sherpa males (though not Sherpa females) were under-represented.

These data raise additional questions. How much do literacy rates reflect local abilities or central inabilities? For example, it has been suggested that teachers are not of a high quality and both teachers and students often play truant. Dropping-out is common. But how important is literacy? In an age of cassette-records, oral cultures may find a new importance and greater relevance to the problems of solving poverty.

Literacy has been linked especially to health, through family planning. Again, the Sacherer (1980:43) evidence shows that Brahmins were more likely to practice contraception than, for example, Sherpas who, in her study, did not practice contraception at all. But the overall level of use was so low that these figures may have no significance. Nor is the evidence on numbers of living children necessarily indicative even though, in one panchayat, it is the 'untouchables,' and lower castes who have the greatest number of living children (5.5) compared to Brahmins (3.3). However, if this discussion is broadened beyond the confines of Sacherer's two panchayats, the conclusions are once more turned on their head. For instance, the military groups (Gurungs, Magars, Rais) and the Sherpas, and a few other groups, have greater exposure to introduced institutions. This is particularly true of the Newars and all the inhabitants of the Kathmandu Valley. The difficulty facing any such discussion as this is not only the unreliability of available data, as has already been emphasized repeatedly, but also the lack of representativeness, so that conclusions based on the detailed study of one or two panchayats may be totally inapplicable elsewhere.

Despite these problems, some inferences are warranted. The tentative suggestion proposed by Pitt ( 1986) is that one contributing cause of poverty is the decline of traditional culture and that introduced institutions, while attempting to work against poverty, may even be exacerbating it. This helps explain phenomena such as relative health without wealth amongst some ethnic groups. In health and nutritional matters, in education, in family life, and so on, some caste and ethnic groups may insulate themselves against outside intrusion. But we must also emphasize Macfarlane's (1976) conclusion which associates increase in population with resource degradation, and hence a breakdown of traditional patterns in all aspects of life, including agriculture, consumption, and so on. As suggested elsewhere, independence and identity are not correlates of poverty, but rather an alternative means of achieving significant life values, wants, and aspirations (Pitt, 1983). Ultimately a firm base of self-reliance may become the best vehicle for promoting economic development. When the independence and identity of a social group is broken down, severe problems may result - not simply the breakdown of social rules and order, but also a continuing dependence, which may or may not be the intention of those newly introduced institutions. At certain times traditional societies are more exposed to intrusions, and perhaps least able to resist, at least in certain sectors. Famines, epidemics, and related disasters, or military disorder, have been the classic examples of outside 'intrusion,' notably on the food or health front. Once involved it seems difficult for outside forces to disengage. The saving grace of Nepal, as opposed to sub-Saharan Africa, may be the ability to prevent the label of disaster being attached so often and so widely. In this context the concept of the preservation of cultural diversity, and hence dignity and self-reliance, must be placed on an equal footing with the ecologists" and environmentalists' demand for big-diversity and preservation of the gene pool. There is a good argument for stipulating that big-diversity and cultural diversity are complementary.