Introduction

As was shown in Chapter One, most refugees attempted to survive without the benefit of the assistance programme. The vast majority of those who were allowed to take decisions themselves, only registered for a settlement under conditions of extreme duress. In 1983 about three-quarters of those interviewed who had settled in the border areas during the first three years of the exodus still maintained that they were 'OK here' (see Table 1:10, page 56). The breakdown of security along the border, and the inadequacies of medical services were the main reasons that people were finally forced to move. The main attraction of the settlements was the availability of clinics and primary schools there. Refugees who arrived during the first three years did instigate the building of many self-help primary schools in the border areas, but the lack of any financial assistance, even for books and stationery, together with occasional government interference,[1] made it impossible to provide enough places for the thousands of children who had been attending school in Uganda.

Other factors from 1982 onwards influenced the movement of people into settlements. The two most immediate obstacles facing refugees after 1982 were the lack of land on which to establish their own farms and reduced opportunities for casual labour in exchange for food. The drought which struck some parts of the district in April/May, 1983, forced large numbers of self-settled refugees to register (at least temporarily) at the Panyume and Kimbe reception centres.[2] Throughout the entire period, but particularly from mid-July 1982, incursions into the Sudan by the Ugandan army influenced some of those who survived these attacks to move to the security of the settlements. But when Sudanese military personnel and certain local officials colluded with the UNLA in 1983, and again, at Kaya, in late 1984, large numbers of refugee households were again disturbed.

In 1983 those refugees who were subject to attacks by the UNLA (and those who were uprooted following the Sudan government's decision to succumb to Uganda's demands that they be removed from the borders) at least had the option of moving to a settlement. But in late 1984 when the problem arose along the Zaire/Uganda border at Kaya, UNHCR had by then determined not to extend its programme by opening new settlements, an(i refugees were left without this option.[3]

Could much of the disruption which families suffered in their country of asylum have been avoided? The evidence suggests that it could. The composition of settlement populations and those who remained self-settled on the border and the economic relations which were established between them, suggest that an assistance policy prepared to respond to need on the basis of demographic changes might have avoided a great deal of the disruption families suffered. Medical and educational services could have been provided. When land ran short, information could have been circulated concerning the whereabouts of other uncultivated areas. Even transport to new places of settlement could have been arranged. But the most important impediment faced by refugees in their efforts to maintain economic independence outside the planned settlements was their lack of security. As one refugee put it. People could have survived well on the borders if, and only if ... the Uganda government [had not been allowed] to violate international law and to come in to harass refugees in another country, another sovereign country ... Somebody who has crossed the border and taken asylum somewhere should be safe. But here to be safe is to be safe in inverted commas. It's unfortunate.'

The deployment

As one UNHCR consultant also discovered, the refugee population is not divided up into two separate groups of settled and self-settled' people (Day-Thompson 1984.) The reality of the situation became apparent to him when he was responsible for moving the thousands of refugees who had been living in the Otogo and Kaya transit camps - some for up to twelve months and more. He came to the conclusion that the refugee population did not comprise one group of 'broken men accepting charity' with the other, the unassisted. constituting 'those brave individualists maintaining their freedom in the face of adversity'. Instead he found that the refugee community is deployed 'like a set of chess-men across the board'.

Believing that refugees simply represent a 'normal cross-section human community, [sic] probably reflecting much of what it was before it had to go into exile', he appreciates their strategies as confirming their capacity to use their 'wits and muscle' to survive in a 'pretty difficult world.'

Refugees, at least some of them, are highly mobile; 45.6 per cent of the self-settled interviewed still lived in the same place as they did on first entering Sudan. Those who remained close enough to the border were able, so long as the opposition remained in control of the area, to slip back into Uganda for food. Up to mid-1982, through binoculars one could see women cultivating on the Uganda side of the Kaya river even though the UNLA was stationed only a few miles away at Oraba. Others were able to slip in and out of Uganda to trade. For example, during 1982-3 all of the fish sold in Yei came from Uganda. Fishmongers travelled through protected corridors from the Kochi and the Nile rivers to the Sudan. They carried receipts proving they had paid the guerillas a tax for this service. Later, when the UNLA made the last push through the north, the fish trade had to be re-routed through Zaire and the supplies in the market declined. Throughout 1982 the town of Moyo was sealed off, with several hundred people living in the mission under guard by the UNLA, but during 1983 trading began between Kajo-Kaji and Moyo. Some refugees particularly women - were able to take advantage of the markets established to serve both sides of the border. A considerable volume of trade also occurs between the Zaire border and Yei. (McGregor in Wilson et al. 1985.)

The refugee chess-board in Eastern Equatoria encompasses not only the settlements, the border areas, and Yei town, but also areas across the Zaire border, Juba - 160 kilometres in the other direction - and distant parts of the Sudan. It also includes Uganda and other African countries. We asked the assisted refugees about the location of their 'relatives' and found that, in addition to relatives deployed in Africa, 180, or 8.9 per cent of the sample, reported having relatives living somewhere outside the African continent.

In addition to these movements and those back and forth into Uganda for trading purposes or to get food, and quite aside from the official repatriation programme mounted by UNHCR, whenever there was a rumour that life was secure enough to do so, many people attempt to return to their homes in Uganda. Refugees in settlements were asked whether members of their household had recently returned to Uganda. They were not asked the reasons for such returns or whether or not they were thought to have returned to stay. The location of relatives according to responses to questions posed to refugees in settlements is shown in Table 3.1.

Fig 3.1. Age distribution of all assisted refugees

Unfortunately, however, one cannot assume with Day-Thompson (1984) that the dispersal of relatives is always the result of strategic planing for survival or that 'relatives' make up a mutual support group. Ugandans distinguish between kinsmen who would normally live in the same compound, the household, and those who are more distantly related by blood and marriage. Norms of reciprocal obligations only obtain between members of the 'household'. Unfortunately only one of the questions posed to assisted settlers about the location of their kinsmen made this distinction. The response to it shows that while people include members of their household when asked about kinsmen, the numbers upon whom they might reasonably expect to depend is far smaller than these answers imply. Although 43.8 per cent of the households in settlements had relatives living in Zaire, only 19.4 per cent had members of their intimate family circle living there. The unassisted refugee households interviewed were only asked one question about the location of kinsmen. They were asked whether any members of their household were living in settlements and 31.1 per cent responded 'yes'.

TABLE 3.1: Location of relatives (assisted)

Although more people simply walk home or 'repatriate themselves' than those who join official repatriation missions, the movement of these people is related to their reasons for flight. Those households which reported having relatives who recently returned to Uganda were analysed in terms of the year they first entered Sudan and by the year the first member of the household came to the settlement. The results are shown in Tables 3.2 and 3.3. These data demonstrate the relative stability, or determined tenacity, of those who both entered the Sudan and came to settlements during the first two years, compared with those who entered later on. None of those who arrived in 1979 reported having relatives who recently returned to Uganda whereas 23.7 per cent of those who arrived in 1982 report at least someone among their kinsmen having gone back to Uganda. These data support the earlier discussion of the causes of exodus for most of the civilians now in the Sudan; there, I suggested that the most important factor in the exodus was the existence of UNLA. It is the undisciplined soldiery of the UNLA rather than opposition to the present government which has led most people to flee Uganda and it is UNLA's continued vendetta against the civilians which keeps them in exile. Had the transitional government in Uganda not foundered and perhaps had the army been integrated on the lines taken after Zimbabwe's independence, or even had the elections been seen to be free and fair, most of the civilians now living in exile would never have fled.

TABLE 3.2: Relatives returned to Uganda by year first entered the Sudan (assisted)

TABLE 3.3: Relatives returned to Uganda by year first household member arrived in settlement (assisted)

Fig 3.2. Percentage absent in each age group by sex

The survey indicates quite clearly that the demographic characteristics of refugee populations vary dramatically. General statements can therefore be misleading. Again, one cannot assume, with Day-Thompson (1984), that all refugees have the same capacity or opportunity to manipulate the advantages of the 'aid system'. All statistical data collected in a social survey must be treated with great caution, but this is particularly the case when studying a refugee population. The analysis of the statistical data included some checks on the internal consistency of informants' responses and some of these will be discussed throughout the book. A reassuring, though very limited, check on the reliability of the data was made by the Oxford team in 1984. They based a more intensive study on the survey sample of households which included if individuals who were 'vulnerable'. They found only seven cases where a year later the same people were not at the same location. However, there was some disparity between the survey data and the estimates of the total population of assisted refugees as made by the UNHCR office. This is not surprising given the difficulties of accounting for numbers under the conditions of an emergency programme and the obvious incentive for refugees to inflate their numbers in order to receive more rations.

In addition to this survey, there were three other sources of population estimates. First, in each settlement the settlement officer was expected to keep a register with the names of family members, and his block leaders also had their lists for food distribution. Second, the UNHCR offices kept an estimate of the total population of settlements and transit camps. This estimate was based on the settlement officer's register and on information on numbers of people who were moved by lorry from borders to transit camps or settlements. Third, the World Food Programme also had a population figure for each settlement, but this estimate was 'negotiated' between UNHCR and WFP to determine the food rations for a settlement. My survey results suggest that of those three sources, the Yei UNHCR office figure was usually the more reliable.

According to the UNHCR office, the total population of assisted refugees at the time the survey was completed was 95,050. Assuming that our sample was representative, the survey found a total population of 106,750 - a surplus of 11,700 (12 per cent). In the notes on methods I describe the efforts we made to ensure that all members of the household were present for the interview. Despite such precautions, one had to recognise that, given the general lack of food, it would be only logical for refugees to attempt to inflate their numbers.[4] Certainly agencies assumed refugees always exaggerated their numbers. We therefore scrutinised those who were absent from the interview because it seemed the obvious place to look to find the evidence for such deliberate inflation. However, when examining the distribution of age groups for the total population compared with the distribution of those who were absent in each age group (as shown in Figure 3.3), it seems very unlikely that in such a large sample informants could have invented an absent population to correspond so perfectly with the age distribution of those who were present for the interview. Moreover, there was no correlation between the degree of correspondence of survey results with UNHCR's population estimates and the numbers absent from the interviews. While one can draw no firm conclusions from the analysis of the absentees, it does reveal the enormous social dislocation which has occurred.

Among those absent were 922 children, (or 39.2 per cent of absentees), who were said to be enrolled in school. In all age groups more men than women were absent, but the differences are not significant. Figure 3.2 shows the numbers and percentage of each age group for both men and women who were absent from the interview. One might suspect that at least some of the men over 15 years were engaged in fighting inside Uganda, but it cannot be assumed all were: on numerous occasions, men who were absent from the interview visited us in our tents in the evening, explaining where they had been.

The population of the transit camps was notoriously unstable, with people moving in and out during the long months of waiting for a settlement. If one accepts UNHCR estimates as reliable, the data from the survey of Goli, Otogo, and Kaya transit camps account for 7,371 of these extra numbers recorded by the survey leaving only 4,329 'extra' to account for. The total population for Otogo transit camp alone was 54 per cent higher than UNHCR estimates. Many people who were in transit camps were unwilling recipients of aid, people who had been forced out of the self-settled areas by the drought. Once the rains resumed, many people returned to their self-settled compounds. Movement in and out of settlements was further encouraged after mid-July when, in the absence of full standard rations, extra rations of edible fat and other items such as blankets were given to refugees so that they could sell these to buy staple foods.

Two transit camps were close to Yei and I met one refugee whose family was in Kaya who considered himself to be part of this camp's population, although he preferred to remain in Yei, unassisted, until a settlement opened. Given the tendency for the self-settled to send some members of their households - usually the more vulnerable - to settlements, informants in transit camps may have included some of their self-settled relatives as 'absent' members of the household. Informants at Otogo transit reported 225 absentees: 61 of these were aged between 10 and 15; 47 men and 49 women aged 15 and older were also absent.

Table 3.4 compares the survey data with UNHCR's population estimate and shows the number and percentage of differences between it and the survey results. It also includes the percentage absent from interview and the average family size for each. The survey found 8 settlements with a smaller total population than that estimated by UNHCR. In 14 settlements the variation (whether higher or lower) between the sample results and UNHCR's estimate was between 0.4 and 11 per cent. There were 5 settlements which had only recently opened - Wonduruba, Dororolili, Dukuni, Adio, and Katigiri. It may be that the progress towards filling Adio and Dororolili had not been regularly recorded in the Yei office, accounting for the higher population according to the survey.

Table 3.4: A comparison of survey and UNHCR population estimates, average household size, and percentage absent at time of interview

Wudabi settlement, according to the survey, had 1,623 more residents than UNHCR believed to be there. Wudabi is only 6 kilometres from the Zaire border and about 40 kilometres by foot from the Ugandan border. Just over a range of high hills to the south is an area where many self-settled refugees live - a half day's walk away. Inside Zaire there is a large 'international' market which attracts refugees engaged in petty trade. It might be assumed that some self-settled refugees, perhaps even including some from the Zaire side of the border, had unofficially moved into this settlement, but they would have had to compete for the rations which were based on UNHCR and WFP lower estimates. It was this settlement in which violence erupted when the supply of food was interrupted. The settlers took up pangas against their settlement officer and accused the local Sudanese officials of having diverted food. Throughout 1983, Wudabi was a difficult settlement to manage.

Wudabi had other attractions besides rations and its accessibility to the Zaire border. It was near a tea plantation which employed a large number of workers, and refugees - mainly children aged 12 and upwards - could be seen congregated at the roadside every morning awaiting transport to work. Just before dark, streams of weary children returned to pick their way across the log bridge over the Yei River and walk home to the settlement. Of those absent from the interviews in Wudabi, 49.3 per cent were between 10 and 19 years of age; 39.9 per cent were below the age of 10. Doubtless some of these children were attending schools, held at that time under the trees.

Even while recognizing that refugees send a disproportionate number of children to the settlements because there are schools available, (see Figure 3.1) and that there is a bias in favour of educating boys, there are still problems in interpreting the data concerning children in the 5 to 9 age group. In 1969, according to the West Nile census, there were 171 children per 1,000 in this age group: 86 boys and 85 girls. According to our survey, there were 236 children per 1,000 in this age group: 127 boys and 109 girls.

Were all the absent children in this age group attending school at the time of the interview? Ugandans put a high premium on educating their children. Early in the interview informants were asked how many children in the households were 'able' to go to school. According to the sample, the total was 36,440. They were then asked how many were actually going to school. The total number reportedly in school was 19,950. As only primary schools (and four intermediate schools) were available for children in settlements, we assumed informants were defining 'able to go to school' in terms of the numbers of 5 to 14 year olds in the household. But later on in the interview, when questions were addressed to each individual, the results show there were 24,200 of the 44,960 children between 5 and 14 years of age who said they were attending school, or 4,250 more children than the number earlier reported as being in school. About half of these were 5 to 9 year olds, and 11,530 were said to be enrolled in school. Of these, 3,210 were absent at the time of the interview. It is difficult to invent the description of an individual required by the interview. These children were born before the civil war began. Their numbers do not therefore reflect either the effects of malnutrition or the hardships of flight on the fertility of their mothers, although both these factors affect children in the under 5 age group who are much more vulnerable to malnutrition and disease. But one might assume that some children in the 5 to 9 age group would also succumb to the perils of flight. A separate examination of even the poor records of deaths available in the settlements showed that large numbers of this age group had indeed died. The nutritional survey carried out in 1984 revealed a very high rate of severe malnourishment among children 5 to 9 years of age (Wilson et al. 1985)

The question of whether absenteeism of this age group reflected unreliability or mobility cannot be finally answered from these data. It is even possible that in some cases the numbers of children in the 5 to 9 age group were inflated by descriptions of dead children. However, one must remember that while the survey of unassisted households did not discriminate comparable age groups, there were only about half as many children in the 6 to 16 age group living in unassisted households.

The deployment of households

It may not be assumed that even members of a household form an interdependent support group. As is true everywhere, the norms of family responsibility as well as the actual behaviour of individual members of the so-called extended family within Ugandan society vary according to the closeness of the relationships as well as by the idiosyncrasies of individuals.

In this study household types were categorized according to the characteristics of the persons registered as the head of the plot in the settlement. 'Normal' families included those where both a husband and a wife were present, and those married women whose polygamous husbands were living in the same settlement. The heads of other household groups in settlements not constituting 'normal' families in these terms were distinguished by age, sex or marital status. These data are compared with responses about the location of their relatives and shown in Table 3.5.

TABLE 3.6 Type of household by size of household (self settled)