SECTOR Home | What's New | About SECTOR | Meetings | Publications | Discussion Forum | Contact Us | Links | Site Map

Agrarian transition in Viet Nam

IV. Incomes and poverty

What sort of income and poverty levels exist in the rural areas, how do they compare to urban areas, and what have been the trends in these magnitudes? These are the questions we shall attempt to answer in this section.

Table 8 provides information on sources of incomes in the rural areas in 1995 and arrives at an estimate of average rural and urban incomes. Agriculture, forestry and fisheries constituted around 63 per cent of total rural GDP in 1995 (compare lines 3 and 5), the rest being incomes from commercial and industrial activities. Agriculture itself generated over one-half of rural incomes, and within agriculture crop cultivation dominated, which in practical terms means paddy. Livestock contributed just over one-fifth of total agricultural GDP. Growth rates shown in the third column point to increasing production in rural areas. A growth rate of 4.8 per cent per annum translates to 2.4 per cent per capita of rural population which cumulated over the decade yields an aggregate increase of around one-third. Equally important was the diversification that occurred in the farm sector as shown by the subsectoral growth rates. Thus livestock production grew much faster than crops, and within the latter, industrial crops increased faster than food crops. To cite some notable examples: output of coffee increased at 33.3 per cent per annum, rubber at 9.9 per cent, coconut 6.7 per cent. Other indicators of diversification will be taken up later.

Table 8. Agricultural GDP, 1995, and growth rates, 1985-95

In the rest of table 8 an attempt is made to derive notional figures of average incomes in rural and urban areas. Much estimation is involved so the results cannot be taken to be anything more than indicative. First an account has to be taken of non-agricultural GDP (industry and commerce) arising in the rural areas. According to estimates provided by the Statistics Office, around a quarter (23 per cent) of the total non-agricultural GDP can be attributed to rural areas. Adding this to the agricultural GDP gives an estimate of rural GDP of 100 trillion dongs. From the rest, which is basically urban GDP, we have to subtract funds devoted to capital formation and add the trade balance to thus arrive at an estimate of urban consumption -- 83 trillion dongs. Divided by their respective populations gives 1.72 million dongs (US$172) as the average rural income and 5.69 million dongs as the urban equivalent. A rural-urban gap of 3.3 is indicated.

Table 9 provides a regional profile of average rural consumption as obtained from the VLSS. First a reconciliation should be attempted with the figures of average income derived in table 8. The years in question differ -- 1995 in table 8 and 1992-93 in table 9. Between these two dates, agricultural incomes in current terms (current terms being the relevant magnitude here) increased by around three-fifths and non-agricultural incomes by four-fifths. If we apply these ratios to the figures of average consumption in table 8, a correspondence is obtained for the rural quantity. The urban 1995-equivalent for the one in table 8 would be 4.3 million dongs, hence short by one-quarter. This figure too is not far out of line with that in table 8.

Table 9. Regional indications of consumption, 1992-95

Great regional variation in average consumption is shown, with the gap between the richest (south-east) and the poorest (north central) region of at least 1:2, even after allowing for price differences. How these figures translate to "real" consumption may be glimpsed from the percentage of income that has to be devoted to food and the calories so obtained. Even in the richest region 53 per cent has to go on food and even then only the bare requirement of 2,250 calories is obtained. In the poorest region the proportion rises to 64 per cent, in other words, only 36 per cent would be available for other basic items such as clothing, shelter, and transport. For most rural households, then, practically all of the income has to go on satisfying their most basic needs. Further conceptualization requires deriving a poverty line and this issue is discussed in the next section.

Income distribution has important rural-urban dimensions. While the mean annual household income in rural areas was VND4.62 million, in urban areas it was VND8.97 million (VLSS, 1994, p. 220). But also between and within rural areas there are wide differences. Two dimensions are shown in table 10. Rural incomes in the southern part of the country far exceeded those in the central and northern regions. Within each region the households that depended on agriculture as the main source of income had the lowest income. Around 55 per cent fell in this category, with a mean income of VND3 million. As compared to this, households that engage in farm and non-farm activities (31.5 per cent) have a mean income of VND5.15 million and households that engage only in non-farm activities (13.4 per cent) VND7.28 million (see table 10). Again income levels vary sharply between regions for each of these categories. In the Central Coast regions farming-only households earn VND1.4 million compared to VND6.9 million earned by mixed households in Eastern South and Mekong Delta regions. These sharp differences arise mainly from differences in location and thus the ecological characteristics of the region and its infrastructure. Rural incomes in the Mekong Delta and the Eastern South regions are high because of good agricultural conditions but also because of non-farm opportunities provided by the proximity to Ho Chi Minh City. In other regions, where farming incomes are lower, so are non-farm incomes. Significant intra-regional income differences also exist and are correlated with average incomes, implying that agricultural development creates stratification. Land distribution currently is relatively even, as, at the end of the collective period, land was allocated on the basis of household size ^(2). However cultivated area is only growing slowly compared to the rural population. In the late 1970s each household had about one hectare of agricultural land, whereas by 1995 this average had fallen to 0.7. Substantial regional differences exist. In the Red River Delta the average is 0.26 hectare against 1.18 in the Mekong Delta. Part of the reason is also unequal access to better quality and irrigated land which allows the better-off farmers to grow more high-value crops and engage in multiple cropping. The richer farmers also apply more purchased inputs in production and engage more in non-farm activities in conjunction with farming (World Bank, 1995b).

Table 10. Annual income of rural households ('000VND)

Rural-urban income distribution has certainly worsened due to the slow growth in agricultural incomes compared to urban incomes. Lagging productivity cannot be blamed for this (tonnage of crops increased at 5.2 per cent annually between 1986 and 1995). Rather, adverse relative prices have contributed. An indirect indicator can be constructed using GDP deflators for the agricultural and non-agricultural sectors. Dividing the two we get a rough "agricultural terms of trade" index (figure 1). These terms improved between 1986 and 1988, but declined thereafter, ending in 1993-95 at a level much below the peak in 1988. An IFPRI study found that the "real" price of paddy (farmgate price deflated by the CPI index) declined at 3.1 per cent per year between 1989 and 1995 (IFPRI, 1996, p. 149). On the other hand, comparing paddy prices with fertilizer prices shows that between 1989 and 1995 the real price of fertilizer declined faster than the counterpart paddy price. However, the greater decline in paddy's real price meant that the profitability of paddy production declined (IFPRI, 1996, p. 329). The main contributory factor here was not trends in the world markets (the export price of Vietnamese rice increased as its quality improved) but the increasing overvaluation of the exchange rate which acted as a tax on export producers (IFPRI, 1996, p. 336) ⁽³⁾.  In addition, significant inefficiencies existed in rice marketing. Rice export is subject to quota, and export licences are only given to state enterprises. The IFPRI study applied a food market model in a number of policy simulations. Removal of the export quota for rice would lead to a significant increase in rice prices and farm incomes. Removal of restrictions on domestic rice trade would lead to more modest improvements.

Figure 1. Agricultural terms of trade

Poverty lines and poverty

The question of poverty in Viet Nam has become controversial with the publication of a World Bank report showing around 50 per cent poverty (World Bank, 1995b). The existence of many different poverty lines also confuses the issue. This being so and the issue important, the accompanying appendix examines the methodology used in the Bank study as well as other studies of poverty in Viet Nam. The conceptualization of the poverty lines is found to be weak and an alternative estimate is presented. Here the main points are summarized.

The most important poverty lines that exist in Viet Nam are:

-- MOLISA's 13 kg per month rice-only poverty line;

-- Nguyen and Nguyen's 16.2 kg per month rice-only poverty line;

-- GSO's food-only poverty line;

-- World Bank's food+non-food and food-only poverty lines.

MOLISA's rice-only poverty line of 13 kg per person per month translates to 1,600 calories per day. Given that the norm is around 2,100 calories, MOLISA's standard is at the barest minimum and is deliberately kept at that level to capture the most abject poor for its Hunger Programme. Nguyen and Nguyen's poverty line would yield 2,000 calories per day and it again is pitched at a basic level. The GSO's food poverty line on the other hand is of quite a different nature. It was derived to counter the World Bank's full poverty line which had yielded the controversial estimate of over 50 per cent poverty in 1992-93. However, the GSO's food poverty estimate -- 20 per cent -- is about the same as the World Bank's estimate -- 25 per cent -- hence nothing is gained. Thus not only controversy but also confusion reigns about the true extent of poverty in Viet Nam and how to measure it.

Basically a poverty line should capture only rock-bottom poverty and for this the poverty bundle should contain only the most basic necessities for survival. In the extreme, a "food poverty line" may be constructed based on the cheapest food item only -- rice in the case of Viet Nam. Anyone falling below such a poverty line could be considered undernourished. MOLISA and Nguyen and Nguyen's poverty lines fall in this category, but not the World Bank's and the GSO's "food poverty lines" since they contain a great proportion of expensive foods. Different "food poverty lines" would obviously yield different estimates of "food poverty" but not all the people so captured would be equally undernourished.

The World Bank's poverty line is the most controversial, giving rise, as we said before, to the alternative derivation of a food poverty line by the GSO. This did not solve the problem since the GSO line was composed of the same sort of foods and thus ended up at the same level as the World Bank's food poverty line. The problem with the Bank's poverty line (whether food or total, since the total poverty line is derived from the food bundle) is that it is based on the food basket consumed by the third richest quintile, i.e. the 40-60 per cent group. This group's food consumption pattern cannot be considered typical of poverty groups -- and is not -- and secondly this benchmark would almost by definition yield an estimate of 50 per cent poverty. In the Bank's poverty basket 75 per cent of calories came from rice, which is of the right order of magnitude, but the rest of the foods, at over four times the price of rice calories, are much too expensive. This means that those who are considered food poor by the World Bank's food poverty line -- i.e. are thought to be obtaining less than 2,100 calories -- would not actually be in that category since they could always resort to less expensive non-rice calories (or in the extreme cut out all of them altogether) to make up their food deficit. From this we conclude (a) a true "food poverty line" should have nothing but rice in it; and (b) if for the sake of variety a more varied diet is stipulated then the non-rice items should be the cheapest ones. This study has postulated two such food poverty lines, one a rice-only one and the other a rice plus some sauce line. A total food and non-food poverty line is also derived. Table 11 shows the basis of these estimates along with those of the World Bank's.

Table 11. A comparison of the World Bank's food poverty line with this study's

Calorie composition (%)			Price per 1,000 calories			Cost of 2,100 calories ('000VND/yr)		Cost components (%)
	Rice	Others		Rice	Others				Rice	Others
World Bank	75	25		505 a	2 391		748		39	61
This study
Rice only	100			482 b			369		100
"First quintile rice+"	75	25		482 b	1 100		488		57	43
Memo items
1. Full poverty line
(a) World Bank							1 090
(b) This study							750
2. Per capita expenditure from VNLSS							1 407
3. Per capita GDP							1 757
a Based on rice prices of D1,784 per kg and World Bank's 3,530 calories per kg which is different from the FAO's figure of 3,700 cals/kg. b 3,700 cals/kg. Source: See Appendix II.

A rice-only bundle could be purchased for half the cost of the food basket stipulated by the World Bank. Admittedly the Bank was not trying to do a rice poverty line, but a comparison against even a more varied "rice+sauce" line also shows that the Bank's "food poverty line" was much too expensive. The fact that 61 per cent of the food expenditure went on non-rice items confirms this, as does a comparison against national average consumption and GDP. The World Bank's food poverty line alone was at over 50 per cent of the average consumption expenditure revealed by the VLSS and its total poverty line at 62 per cent of per capita GDP.

An internal calculation shows that five per cent of the population fall below the rice-only poverty line. In practical terms a person at that kind of income threshold (369,000 dongs in 1992-93) would be consuming at the most 1,700 calories per day to leave him some surplus for basic non-food essentials. Just around one-quarter of the population falls below the full poverty line, the same figure as obtained for some kind of "food poverty" in the World Bank document -- the same, since our full poverty line and the Bank's food poverty line coincide. In practical terms this kind of "poor" people should be otaining 2,100 calories from a tolerable diet and then having a surplus of another 262,000 dongs for non-food needs. Regional figures for total poverty would also be the same as for food poverty in the Bank report. These estimates are shown in table 12 along with figures for average consumption as obtained in the VLSS.

Table 12. Poverty estimates based on full poverty line,

and per capita expenditure

Plotting the two sets together would show a clear inverse correlation between poverty and income: poverty declines as we move up the income ladder regionally. Certain characteristics of the poor compared to the rich are shown in table 13. The poor have larger households and more children than the rich; they have lower literacy rates and fewer years of schooling; they have limited access to safe water, sanitation and electricity; their children suffer more from malnutrition, wasting and stunting, and they are more prone to disease.

Table 13. Characteristics of the poorest and richest households, 1992-93

Wide regional differences exist, with educational levels being lower in the Northern and Central Highlands than in the Mekong Delta. Related to the regional pattern is the lower level of attainment for ethnic minority groups (Nguyen Thi Canh, 1996). Educational attainment does not vary by sex however (with the exception of the ethnic minorities where female participation in education is lower). Other characteristics of the poor also appear from the VLSS data (VLSS, 1994; World Bank, 1995b). Most of the poor (90 per cent) live in rural areas and are predominantly engaged in agricultural activities; of the households whose head is occupied in agriculture, 30 per cent fall below the poverty line (the highest) compared to 10 per cent for white-collar workers (the lowest). Poor households tend to have smaller landholdings but larger families, with women having a higher fertility rate.

As for trends in rural poverty, all indicators show that poverty levels have declined continuously since the mid-1980s. Growth rates experienced in the rural sector shown in table 8 attested to this as do figures presented in table 14 for paddy, paddy being an appropriate proxy to use since it still constitutes an important part of rural incomes.

Table 14. Paddy production per capita, 1976-80 average, 1980,

and 1985-95 (kg)

Paddy production increased by 55 per cent in per capita terms between 1980 and 1995, which translates to 64 per cent per capita of farm population. As shown in Appendix III, the 1995 paddy crop would supply 1,927 calories per person per day, due account being taken of output used for seed and , feed and waste, and output exported. On a similar basis the 1980 paddy production would translate to 1,245 calories. Since the output of other foods was also much lower then, food supply in 1980 was probably no higher than 2,000 calories per capita, compared to 2,463 for 1995 (see table A.III.1), or below the norm of 2,100 calories. Thus considerable improvement in dietary standards is shown for 1995 compared to 1980. Food-related indicators reflect this improvement. Life expectancy at birth increased from 63 years in 1980 to 68 in 1994 (World Bank, 1997) while the infant mortality rate (per 1,000 live births) declined from 82 in 1977-83 to 55 ten years later (ICDC, 1996).

Although this paper is about the rural sector some comments on urban poverty are in order since the two types of poverty and living standards are ultimately linked. Compared to the estimate of 25 per cent poverty in rural areas in 1992-93, in the urban areas the corresponding figure is 10 per cent. Such a low level of poverty should be considered remarkable since just before this time employment in the state sector had declined drastically and wages had fallen. The wage story with all its ramifications is taken up with the help of figure 2.

Figure 2. Real wages of government workers, 1985-97

Already by 1991 real wages of both civil servants and state enterprise employees had fallen by one-half compared with 1985. Considering that the 1985 wage was barely above the sufficiency level, the post-reform declines meant increasing hardship for workers. The 1989 government average wage (D38,045 per month) would purchase only 2.3 kg of rice (price D550 per kg) per day and the state-enterprise wage (D54,522) 3.3 kg. At mid-1995 the purchasing power of the average government wage (D300,000 per month) had climbed to 3.3 kg of rice (rice price D3,000 per kg) and by the end of 1997 to 5.3 kg. Thus the average wage has recovered to its pre-reform level, and although only US$40 per month, is not as dire as it may seem; in Appendix II a rice+sauce poverty line is calculated at just under 2,000 dongs per person per day at December 1997 prices, or D3,5000 allowing for non-food needs. The average wage at D400,000 per month would thus support just under four persons. Some supplementary income would be needed to support an average family of five persons and this has been forthcoming in the form of growth of the informal sector, so that altogether the average urban family has improved its position considerably, particularly since the start of the 1990s decade.

All rural communities have primary schools, but the facilities are meagre. Lack of teaching materials is a common complaint and the low salaries ensure that the teachers are only available for teaching part time (World Bank, 1996). With the demise of the cooperatives the collective care for rural education and health fell away and was replaced by private contributions. According to the VLSS, educational expenditure per primary-school student in rural areas was VND54,900 per year in 1992-93 (VLSS, 1994, p. 73). This is a sizeable sum in comparison to the mean annual rural household income of the lowest income group (VND2.59 million) and it should then be no wonder that the poor drop out of school. The net enrolment ratio was calculated at 68 per cent for the lowest income quintile and 86 per cent for the highest (World Bank, 1995b, p. 84). Of those not attending school, 44 per cent cited cost (World Bank, 1995b, p. 201). As state expenditure on education strongly concentrates on post-primary education it mostly benefits the rich, since poor children have dropped out by then.

The rural health system too suffered from decollectivization. While almost all rural communes still have at least one clinic and a physician, the facilities are invariably rudimentary (World Bank, 1995b, p. 213). As with education, the demise of the cooperatives has meant a reduction in public funds for equipment and medical supplies. The solution here also is a resort to user fees and private doctors. Average expenditure on health in rural areas varies from VND44,100 for the lowest income group to VND228,410 for the richest (VLSS, 1994, p. 111). The increasing privatization of health costs takes its toll in terms of reduced participation of the poor. The VLSS reported that only 14.5 per cent of those in the poorest income quintile consulted medical services in the month before the survey, compared to 27.2 per cent in the highest quintile. The utilization rate of health services (public and private, including hospitals, doctors, paramedics, but excluding self-medication) was only 0.79 contacts per person per year for the poorest group compared to 1.52 for the highest (World Bank, 1995b, p. 205). Such trends in the use of social services quite evidently constrain the possibilities for the poor to escape poverty. Studies show consistently that educational attainment and health status are important contributors to agricultural productivity (World Bank, 1995b, Annex 3.2). The increasing cost of education and healthcare and the consequent lower participation of the poor, who are predominantly rural, constrains agricultural growth and risks creating a class of underprivileged with little chance of escaping poverty.

Also it is worth noting that while incomes of the poverty groups -- farmers and wage-earners -- have increased, they have not done so by anything like the growth in average urban incomes (over 10 per cent per annum between 1985 and 1995) so that to that extent income distribution has worsened, both along rural-urban and intra-urban lines. A mitigating factor is that income that did not go to the farmers or wage-earners which would normally be called "profit" has gone to the Government in the form of surpluses. Much of this has been utilized to maintain the level of social services in the country.

International Labour Organization (ILO): Contact us | Site map |

Copyright and Permissions 1996-2008 International Labour Organization (ILO) - Disclaimer