Violence and insecurity: 1980s to mid 2000s

Violence associated with cattle raiding became a pervasive threat in Karamoja starting in the 1980s.Footnote ⁵ The raids of the 1980s, 1990s and early 2000s were normally large and well-organised events. They most often occurred in locations where the raiders could expect to find a large concentration of animals – such as kraals – and attackers were usually heavily armed. Dyson-Hudson (Reference Dyson-Hudson1966) describes individual Karimojong raids in the 1950s as involving up to 600 men and resulting in the losses of several thousand head of cattle. Young men were the most likely to experience casualties, as they were at the front lines of both raiding and defending cattle (Gray et al. Reference Gray, Sundal, Wiebusch, Little, Leslie and Pike2003). Many authors argue that strict codes of cattle raiding enforced by male elders ensured that women and children were spared from the most egregious violence (Adan & Pkalya Reference Adan and Pkalya2005; Simala & Amutabi Reference Simala, Amutabi, van Schendel and Abraham2005; Akabwai & Ateyo Reference Akabwai and Ateyo2007; Carr Reference Carr2008). However, such prohibitions were not always effective, and women and children were increasingly killed or injured in raids. In a 2009 group interview in Kotido District, female respondents explained that although men made up most of the residents at kraals – and hence were the most exposed when attacks took place – women and children could also be in great danger if found at kraals and could sometimes be subject to targeted attacks.

Somewhat counter-intuitively, women and children often had more protection during attacks on the settled manyattas, as most of the fighting took place outside. The same group of women in Kotido explained, ‘fighting mostly affected men – they were the ones outside the home. The women were in danger only if we came outside to relieve ourselves and found that the enemy was hiding outside. Then you would be killed.’Footnote ⁶ Extreme violence and widespread casualties often occurred during revenge attacks (Eaton Reference Eaton2008).Footnote ⁷ A Voice of America report from September 1999 captures the scale of a particularly large attack:

A reported increase in sexual violence in the 1980s and 1990s may also have been the result of revenge mentality.Footnote ⁸ A group of Matheniko male elders interviewed in 2008 reported, ‘In retaliation, the avengers go for women, children, the lame and elderly to pay for their lost wives or sisters! That is why even the raping came in. It was a tit-for-tat affair as raiders would rape any women they would find on their way in revenge for what was done to their manyatta or kraal.’Footnote ⁹ However, raiders from Karamoja were also using sexual violence to terrorise the largely unarmed populations in the neighbouring Acholi and Lango regions during this time (Stites et al. Reference Stites, Mazurana and Carlson2006), and hence sexual violence cannot be explained solely by revenge.

These examples provide a window into the insecurity in the region during this time. Not all experiences of violence were so extreme, and those that were most extreme were more likely to make headlines or to remain in people's memories. However, the threat of violence was ubiquitous and limited residents' access to grazing areas, markets, natural resources and social networks.

The rise and role of commercial raiding

In Karamoja, commercialisation was an important factor in the rise of violent raiding in the 1980s and 1990s and a component of the gradual transformation of raids from large communal activities to small group or individual endeavours aimed in part at meeting needs for food, commodities and other expenses. Commercial raiding here refers to the exchange of stolen animals for cash or other commercial gain.

While commercial raiding is tied to the evolution of pastoral societies in East Africa, there is no clear date when this practice began. Variations exist based on location and historical interpretation. For instance, in examining the Kuria of Tanzania, Fleischer (Reference Fleisher1998, Reference Fleisher2000b) argues that raiding evolved from a form of cultural expression to a capitalist strategy beginning in the colonial era. Anderson (Reference Anderson1986) has similar findings about the colonial roots of commercial raiding among the Kalenjin in Kenya. Some scholars of Turkana (Kenya) and Karamoja locate the emergence of commercial raiding in the decades following national independence (Ocan Reference Ocan1992; Mirzeler & Young Reference Mirzeler and Young2000; Markakis Reference Markakis2004). Mkutu (Reference Mkutu2003), on the other hand, believes that the commercial component did not become the main motivation in Karamoja until the mid-1990s. Regardless of the time period and region, the commercialisation of raiding is best understood as a gradual and irregular process: some raided cattle in some areas are sold for commercial gain at certain times, while on other occasions the same raiders may retain or exchange stolen animals for traditional purposes, such as marriage, tribute and increasing one's own herd size.Footnote ¹⁰

Raiding parties decreased in size as raiding transformed in East African societies experiencing commercialisation. By and large, commercial raids in the Greater Horn did not have the approval and blessing of the male elders and community members that had existed for traditional raids (Mirzeler & Young Reference Mirzeler and Young2000; Akabwai & Ateyo Reference Akabwai and Ateyo2007). Young men engaged in raids that took place outside of official sanction sought to avoid detection, and hence were more likely to operate in small groups. Anderson traces the transition of stock theft from a communal activity mediated by custom and collective justice to a form of ‘sport’ for groups of young Kalenjin men in Kenya (Reference Anderson1986). The shift from the large group to the individual (or small group) undermined the authority and control of local elders who had previously been involved in planning raids. It also led to increased militarisation as more guns came into the region and were purchased at low prices by young men (Mirzeler & Young Reference Mirzeler and Young2000). Importantly, Markakis points out that, even when it was a communal activity sanctioned by elders, raiding served as a means through which ‘young men acquire their own herds and assert their manhood’ (Reference Markakis2004: 26). The role of raids in asserting manhood became even more important in Karamoja as traditional rites of passage for male youth (namely initiation and official marriage) became less common (Stites Reference Stites, Tripp, Ferree and Ewig2013b).

Shared benefits from raids decreased as raids shifted from the communal to the commercial and raiding parties became smaller. Animals from successful traditional raids had been incorporated into collective herds and paid as tributes to those who had planned and blessed the raids, including the elders. Commercial raiding, in contrast, largely benefitted the raiders alone, with fewer benefits (or blessings) passed vertically. Horizontal transfers that did continue were more circumscribed, with benefits going to the raider's friends, his immediate family, or to an in-law in the form of bridewealth.

Reflecting on periods of intense raiding in the 1980s to early 2000s, respondents in Karamoja often cite the involvement of businessmen, politicians, police and members of the military. While difficult to prove, these rumours of outside involvement are notable because they are so widely believed by the local population. Many sources on commercial raiding in East Africa explore the role of external actors.Footnote ¹¹ This ‘external collaboration and assistance’ (Oloka-Onyango et al. Reference Oloka-Onyango, Zie and Muhereza1993: 12) can reportedly take a variety of forms, such as an extended entrepreneurial chain originating with a specific order for cattle (Schilling et al. Reference Schilling, Akuno, Scheffran and Weinzierl2011), training in military tactics and use of modern weapons (Ocan Reference Ocan1994), the involvement of external ‘armed military or bandit groups’ seeking to ‘procure cattle in vast quantities either to feed warring armies or to sell on the market’ (Hendrickson et al. Reference Hendrickson, Armon and Mearns1998: 191), logistics and transport support (Akabwai & Ateyo Reference Akabwai and Ateyo2007), and the financing of cattle raids for commercial purposes (Mkutu Reference Mkutu2003).

Our data show that networks of raiding and commercial exchange in Karamoja are broader than the immediate actors or nearest market. Overall, however, these data illustrate a broader and less nefarious view of the commercialisation of raiding in Karamoja from that depicted above. This is in line with Krätli & Swift who point out that

In other words, the transfer of stolen animals is commercial whenever such animals are exchanged for cash or items in kind, as opposed to being used to build herds, for a ritual or social event (sacrifice, marriage, etc.), or for a reciprocal social transfer (repayment of debt to a stock associate, lending to a relative, etc.). Our analysis shows that this more nuanced interpretation of commercial raiding allows for an analysis of individual motivations and smaller scale transactions, and helps explain the prevalence of individual and small-scale thefts seen today.

The role of poverty in commercial raids receives less attention than the shadow markets and the cross-border trade in small arms, but is central to understanding past and present patterns of violence. Writing on Tanzania, Fleisher and Holloway (Reference Fleisher and Holloway2004) point to the greater likelihood of raiding by males from poorer households. This was especially the case for men who had no sisters and therefore lacked the promise of bridewealth animals from the sisters’ marriages. More broadly, the lack of economic development in most pastoral areas means there are limited opportunities for income generation for young men outside of the pastoral economy (Catley & Aklilu Reference Catley, Aklilu, Catley, Lind and Scoones2012). Fleisher states: ‘Kuria cattle raiding today is most productively viewed as one item in a desperately brief menu of cash-generating economic options available to the Kuria people of rural Tanzania’ (Fleisher Reference Fleisher1999: 240). Krätli & Swift (Reference Krätli and Swift2003) make the distinction between men who engage regularly in raids as a way to make a living and those who undertake periodic raids to restock, feed their families or acquire money for school fees. The causal link between poverty and raids is highlighted by Gray et al. (Reference Gray, Sundal, Wiebusch, Little, Leslie and Pike2003) who argue that raiding is an adaptive cultural response (albeit negative and destructive) to environmental, sociopolitical and economic pressures upon pastoralists’ subsistence base. This economic motivation is strongly supported by our research on Karamoja and underpins the shift from raids to thefts in the region, to which we now turn.

Shifts from ‘raids’ to ‘thefts’

Our research over the past decade in Karamoja illustrates a change in how attacks are described, with victims making a clear distinction between ‘raids’ and ‘thefts’. In addition, a new type of thieves, considered ‘thugs’ and widely condemned, has emerged. Thefts have become increasingly opportunistic and spontaneous (as opposed to well-planned raids), and are carried out individually or by a handful of men. Perpetrators are either unarmed (although they may pretend to be), carrying knives or sticks, or sharing a weapon among a group. Raiders of the past were interested primarily in livestock, with an over-riding preference for cattle. In contrast, the thieves of recent years are more likely to steal productive assets (such as tools), essential domestic items (clothing, bedding and cooking implements), and food. Livestock may be stolen if the opportunity arises, but various factors make this a risky venture, including the increased efficacy of tracking systems and punishments. Productive and household commodities, in contrast to livestock, can be quickly converted to cash on the second-hand market or incorporated into thieves’ households.

The desire for cash that predominated in the commercial raids in the final decades of the 20th century is a central component of the thefts of today. When asked why men steal, the most common response in all locations and by both genders was ‘hunger’. The lack of income-generating alternatives for young men without animal herds is seen as a driving force behind the thefts. Poverty is not the only factor, but understanding the connection between asset theft and basic needs explains at least some of the motives for such crimes. Unlike the raids of the past, when proceeds were widely shared, looted proceeds today are largely kept by the individual or his household.

In 2008, reports began to emerge of thieves or thugs who attacked homes and homesteads on a regular basis (Stites & Akabwai Reference Stites and Akabwai2009). At the time these thieves were called lonetia, a reference to the widespread theft of household goods and grains that were stored in mosquito nets.Footnote ¹² The lonetia attacks were notable for many reasons, including the shift to private residences, which brought an expansion in types of victims. A group of young men interviewed in Kotido in 2009 reported:

Sexual violence appears to have increased in parallel to attackers entering domestic spaces (Stites & Akabwai Reference Stites and Akabwai2009). Such violence also increasingly took place upon women collecting natural resources in the bush, as described by young female respondents in Moroto in 2009, ‘When you are attacked in the bush they take your clothes, your beads, your sheets – anything of value. You might also be raped and even get ‘slim’ [HIV/AIDS]. Rape is not that common unless it is very bad person.’Footnote ¹⁴

The strategy of opportunistic theft (with or without accompanying violence) was applied regularly if infrequently by young men who felt they had few options. A longitudinal quantitative study with approximately 400 young men in southern Karamoja in 2013 found that many who stole did so only once over the four time periods of study. These once-off actions followed a seasonal trend and increased during the hunger gap. A much smaller subset of young men engaged in such theft regularly and as a primary livelihood activity. Importantly and in marked contrast to the previous insecurity in which attacks were almost exclusively inter-group (or inter-sub-county, in the case of raiding among the Pian), lonetia at times steal from within their own communities, a previously unknown phenomenon (Stites & Marshak Reference Stites and Marshak2016).

Lonetia attacks were a consistent problem in most study sites and were a major security concern for respondents during fieldwork from 2008 to 2010. This corresponded to a time in which disarmament abuses by the military were widespread, animal deaths in protected kraals Footnote ¹⁵ were high, and skirmishes between groups remained rampant. However, perceptions on overall levels of security shifted markedly over a period of several years, even while lonetia activity continued. The thefts still brought condemnation from community members, but the exacerbating conditions had improved. By 2013, many respondents deemed the insecurity of these small-scale thefts to be minor in comparison to previous experiences of violence and upheaval. This sentiment is captured in the thoughts of male youth respondents in Nakapiripirit in April 2013, ‘The people can walk to other villages without fear. Women can go to fetch firewood and water freely. They visit each other. People can now sleep at night peacefully. The only problem now is lonetia.’Footnote ¹⁶ By 2015–16, respondents in many areas only mentioned lonetia thefts when asked directly, and depicted the incidents as a common nuisance to be dealt with as best as possible. We posit that while these improvements are part of an overall decrease in insecurity and violence, low-level theft remains pervasive and has the potential to once again escalate into a wider problem.

Impacts of improved security

Reductions in violent, large-scale and unpredictable attacks have brought a number of positive impacts as reported by both men and women in multiple locations. One of the most critical impacts is improved mobility for both humans and animals. For example, in 2009, women's fear of attack when collecting natural resources was so pronounced that some communities in southern Karamoja were dismantling their manyatta fences to use for firewood instead of gathering wood in the bush. In contrast, in 2015 and 2016, women in both northern and southern Karamoja reported that they move freely to collect resources, to markets, and even to neighbouring districts to buy sorghum or other commodities.Footnote ¹⁷

Animal migration patterns, once severely constrained by insecurity, have recovered to an extent not seen in decades. Male herders interviewed in western Kotido District in 2016 explained that in previous years they would have kept animals of different types in one large herd for ease of protection, despite related disadvantages for animal nutrition and health. They reported, ‘Now that it is peaceful we don't need to have them all in one big kraal like that. We are only able to do this dividing by type because of the relative peace.’Footnote ¹⁸ This disbursement allows for different types of animals to access the specific and most appropriate rangeland types.

Relations among groups have also improved, allowing for increased dry-season access to water and grazing points. Animal owners and traders move easily between districts within Karamoja, including on foot and accompanied by animals, to buy and sell at the main markets.Footnote ¹⁹ Dry-season grazing is now expansive and includes sharing of pasture among groups. As of early March 2016, Jie herds from Kotido were in Abim District, Pokot herds from Amudat were in Kween (formerly Kapchorwa) District, and Pian herds from Nakapiripirit were in Napak District. Matheniko and Jie herds were soon to join up with Turkana animals on the eastern edge of the region. Historical precedents allowed access to all of these areas, but insecurity had undermined usage rights for many years. While isolated problems continue (such as between Acholi and Jie populations on the Kotido–Agago border at the time of our visit), these issues are minor compared to the insecurity that previously made these areas entirely off-limits.

Insecurity from livestock raids is still a concern in some locations, particularly cross-border raids from Turkana and Toposa in Kenya and South Sudan respectively. Security personnel remain actively involved with kraals in select areas, such as along the Uganda–Kenya border in Kaabong District. Kraals in this area remain under UPDF and LDU protection, but even these arrangements are continuously evolving as security improves. For example, although a large protected kraal (called Losera) still existed in March 2016 in Kalapata Sub-County in Kaabong, several smaller kraals had broken away from the main one with the permission of the UPDF. These kraals have more mobility and are less crowded than Losera, but are still guarded by LDUs. Similarly, a Pokot kraal in Kween District was fully mobile and yet had LDU protection. The willingness of the security forces to allow such innovations – and, in the case of the Pokot kraal, to move far distances with herds – is an example of the security forces’ increased adaptability, improved community interaction, and better understanding of the needs of pastoral populations. These differences are particularly stark when compared with the early phase of the 2006 disarmament campaign and the establishment of the protected kraals (Stites & Akabwai Reference Stites and Akabwai2010).

Factors of change

Many of the improvements to security detailed above have been gradual and non-linear, meaning that outbursts of violence or a spate of raids may have occurred even while the overall trend was in a positive direction. One of the clearest examples of this uneven trajectory came in the early years of the disarmament campaign that began in 2006. Many communities experienced an initial increase in insecurity due to uneven disarmament and abuses by the armed forces (Stites & Akabwai Reference Stites and Akabwai2010). Overall, it is difficult to quantify shifts in violence or crime in Karamoja due to the unreliability of police reports, death certificates, and hospital records. This article does not claim that the changes in violence and security are absolute or immutable, but data from repeated research trips to multiple locations within the region provide strong and compelling evidence for overall improvements across broad swathes of Karamoja.

The overall and marked positive changes in security are the result of several inter-related factors. The first is the now 10-year-old disarmament campaign. Although beset with early problems (Human Rights Watch 2007) and – like previous campaigns (Knighton Reference Knighton2003) – extremely unpopular, local perceptions began to change in approximately 2012 (with variations based on location). Today, male and female residents in both rural and urban areas express overall support for the outcomes of disarmament. This is the case even in some areas that had markedly negative experiences with the UPDF and the disarmament process. In one such example, a group of young Jie men interviewed in early 2016 reported that peace had been delivered by ‘a good leader [President Museveni] who removed the gun’. We asked about the discrepancy between this view and that expressed in the same location in 2009, a time at which young men in particular considered the UPDF a mortal enemy and helicopter gunships had attacked herders who resisted disarmament in a nearby sub-county. The young men interviewed in 2016 responded, ‘When [the UPDF] first came here it was very hard and there was a lot of attacking. But now we see the better way of living. It is like when you beat a small child who corrects himself and grows into a good adult.’Footnote ²⁰ This could be interpreted to mean that – with hindsight – the respondents appreciated this ‘corrective’ external use of force, or, on the contrary, that they realised that fighting against the authority of the state was a lost cause.

As a corollary of the success of disarmament, the UPDF as an institution is today largely viewed in positive terms. Male respondents in Kotido and Kaabong in early 2015 reported positive cooperation between the UPDF and local peace committees in tracking stolen cattle. Female respondents in manyattas near UPDF barracks also had favourable views, and reported that the soldiers often share food and health resources.Footnote ²¹ Importantly, these impressions were in contrast to views regarding the police, who were seen as either absent or unwilling to assist residents without a bribe. These negative and suspicious impressions of the police existed in most locations we visited in 2015 and 2016.

A second factor contributing to improved security is the increased involvement of community-based and cross-sectoral mechanisms in combating crime. The most widely cited of these bottom-up processes is the ‘two-for-one’ policy, whereby an apprehended thief is required to hand over two times the number of animals he stole, plus one additional animal. This initiative arose out of a meeting at Nabilatuk, Nakapiripirit District in April 2013 attended by representatives from local communities, the police, UPDF, and national and international conflict mitigation actors. The Danish Refugee Council/Danish Demining Group (DRC/DDG) and other external actors facilitated the meeting, but the resolution was reached by participants over the course of three days of discussions.Footnote ²² Often referred to as the Nabilatuk Resolution in southern Karamoja, communities in northern Karamoja agreed to a similar mechanism known as the Moruitit Resolution. In response to these initiatives, communities formed peace committees (consisting of male elders, male youth, and members of the UPDF and police) at the local, sub-county and district levels. These volunteer committees enforce the resolutions and coordinate the tracking and recovery of lost animals among the local volunteers and the security forces. Respondents in multiple sites in 2015 and 2016 cited the extent of local acceptance of and adherence to these resolutions, as well as the effectiveness of the ‘two-for-one’ policies and the peace committees in improving law and order.

These local mechanisms are not without their problems. Recent data reveal that the terms of Nabilatuk/Moruitit policies can be implemented with some discrepancy. For example, if the perpetrator is ‘too poor’ he can be excused from the requirement to hand over twice the number of stolen animals. The animals actually stolen must still be returned, and perpetrators unable to return the animals are taken to the police.Footnote ²³ Local communities feel largely positively about these systems, but some point to potential injustice in the collective retribution that stipulates a return of double the looted animals. For instance, if an alleged thief does not own a sufficient quantity of animals to pay the two-for-one fine, then the animals are meant to be repaid from other sources. As explained by the UPDF Brigade Commander of Kaabong, ‘When you have to pay back what you don't have, you need to go to your relatives. And if the relatives don't have, then it is the community that must pay.’Footnote ²⁴ Those within the community who have few livestock stand to suffer disproportionately in the absence of local safeguards.Footnote ²⁵ Another criticism has to do with the extent of involvement by external actors in ‘imposing’ these peace mechanisms as part of the ‘imposed pacification’ of the region (Caravani Reference Caravani2018). At this juncture it is difficult to distinguish between ‘facilitation’ of the initial Nabilatuk meeting and a more heavy-handed initial involvement. It should be noted, however, many local respondents spoke proudly about the local and grassroots nature of these resolutions.

A third and related factor promoting better security in the region is what many respondents described as the revitalised authority of the elders. In most locations visited between 2013–16, both male youth and male elders described improved inter-generational relations compared with 5–10 years earlier.Footnote ²⁶ Youth felt that elders were approachable for both advice and support, and youth felt included in decision-making.Footnote ²⁷ Elders largely agreed with these depictions. This positive picture is in contrast to that of the decades prior to the 2006 disarmament in which violence had increased and elders viewed youth as disrespectful, stubborn and disobedient. Elders largely attributed this disharmony to ‘the gun’; many elders had wanted male youth to cease raiding that was seen as individually driven and without communal benefit, but these efforts were largely to no avail (Akabwai & Ateyo Reference Akabwai and Ateyo2007).Footnote ²⁸ Female respondents also perceived and benefitted from the improvement in male intergenerational relations, as illustrated by a woman interviewed in 2016:

A lack of power-sharing underscored much of the tension between generations of men. Initiations, particularly in southern Karamoja (among the Bokora, Matheniko and Pian) had stalled due to a delay in the transfer of authority from one generation to the next as stipulated through the gerontocratic generation-set system common to pastoral groups in East Africa (Gulliver Reference Gulliver1953; Spencer Reference Spencer1976; Knighton Reference Knighton2005).Footnote ³⁰ Insecurity, poverty, drought and reluctance on the part of the senior elders to acquiesce all contributed to the failure to hand over power and to begin initiations of a new generation of men (Stites Reference Stites, Tripp, Ferree and Ewig2013b).Footnote ³¹ Young men were not able to initiate in the absence of this transfer, and hence lacked official adult status regardless of chronological age. (Many respondents were in their late 30s and even 40s and had still not initiated.) The lack of transfer of power and the stagnation of initiations further fuelled the discord between generations of men in the late 20th century, and many observers felt that this contributed to violence in the region (Gray Reference Gray2000; Stites Reference Stites, Tripp, Ferree and Ewig2013b).

Succession began across much of southern Karamoja in approximately 2013, and initiations soon followed (Gray, personal communication, February 2014).Footnote ³² The renewal of initiations has directly improved the relationship between male youth and their elders: the youth are reportedly affording greater respect to elders who, in turn, are seen to be acting as the appropriate guardians of power.Footnote ³³

Arising in part from the improved relations, male elders in some areas report being able to revitalise customary justice mechanisms that the youth had previously either largely ignored or deemed irrelevant. One such mechanism is ameto, through which male elders order the caning of a wrong-doer by his peers along with the payment of a fine (usually a bull). The threat of ameto reportedly helps to keep male youth in line (Muhereza et al. Reference Muhereza, Ossiya and Ovonji-Odida2008; Gray Reference Gray2010), and appears to have a direct impact on improved security (Howe et al. Reference Howe, Stites, Akabwai and Corps2015; Stites & Marshak Reference Stites and Marshak2016).Footnote ³⁴ However, there are limits to its effectiveness as a deterrent, including that it can reportedly only be applied to those of the same clan.Footnote ³⁵

Many respondents drew a connection between the revitalised power of the male elders and the two-for-one policies in response to animal theft. The possibility of losing animals from the community to compensate for the crimes of recalcitrant youth may have helped to galvanise the male elders. In addition, the coming to power of the next generation of elders as part of the recent succession has increased their legitimacy in the eyes of the youth and has helped to improve vertical relationships.

Respondents interviewed in all sites in 2015 and 2016 felt that security had greatly improved, but widely described the status quo as one of ‘relative peace’. Those interviewed emphasised that peace is fragile given the continued presence of weapons in neighbouring Kenya and South Sudan. Additional factors of concern voiced by key informants include potential political upheaval at the national level or a too-rapid reduction of UPDF forces in the region.Footnote ³⁶ A general lack of trust in the capabilities and competencies of the police compound these concerns. At the local level, respondents qualify their views on the current peace due to the continued patterns of theft discussed earlier and the pervasive domestic violence to which we now turn.

Domestic violence

Sexual and gender-based violence and, in particular, domestic violence against women by male partners,Footnote ³⁷ is certainly not new in Karamoja. Efforts to measure or quantify the extent of the problem or changes over time are unreliable due to the sensitivity of the issue. Its long-standing existence does not, however, make its pervasiveness or impacts less of a problem, and the seriousness is increasingly being recognised in official circles. In 2015 and 2016, for instance, sequential District Police Commissioners (DPC) in Kotido reported to the research team that domestic abuse was one of the most pressing concerns facing the district.Footnote ³⁸

In interviews in Kotido and Kaabong districts in early 2015, more than half of the women interviewed in focus groups in 13 villages (across 11 sub-counties) reported being victims of domestic violence. Domestic abuse is seen as nearly constant, as indicated by the following quotation:

Many of the female respondents reporting abuse had suffered serious injuries, including blindness, broken bones, burns and deep cuts. Several had permanent disabilities resulting from the violence. These acts were at times torturous in nature, such as kicking a pregnant woman in the abdomen, burning of genitals, gouging of eyes and cutting of throats. Death was not uncommon: in one sub-county of Kaabong, five women had been killed by their husbands in January 2015 alone (Howe et al. Reference Howe, Stites, Akabwai and Corps2015).

Male respondents interviewed in northern Karamoja in early 2015 were generally open about domestic abuse. Men considered violence to be justified if a woman had not prepared or had burned food, was ‘lazy’, came home late, ‘did not take care of the children’ (as evidenced by children crying or misbehaving), challenged authority or acted disrespectfully, drank to excess or committed adultery. In contrast, very few female respondents believed that violence perpetrated by a husband was ever justified.Footnote ⁴⁰ Women reported that refusing to have sex was the most frequent ‘cause’ of domestic abuse. Without exception, women said that they had no right to refuse sex. Violence within the home most often occurs when the perpetrator is intoxicated (Howe et al. Reference Howe, Stites, Akabwai and Corps2015).

Domestic abuse in Karamoja is both a security and an economic issue. Women in Karamoja provide for their households largely through the production or sale of natural resources, through cultivation, or through service-sector work in towns. The extent of women's economic contributions to their households is particularly pronounced due to the decrease in traditionally male-dominated animal husbandry. Due to the extent of injuries from domestic violence, women are at times unable to work to support their households for days, weeks or months at a time.

Perceptions on whether domestic abuse is increasing or decreasing differed by location, and potential explanations exist for both trends. A shift in gender roles within households is one possible cause of increased violence. The decline in animal-based livelihoods due to growing inequity of livestock ownership and the protected kraals as part of disarmament removed many men from their herds. Whereas previously these men spent long periods of time in the kraals, they are now more likely to be home and with few meaningful tasks to occupy themselves. Women, on the other hand, are increasingly the main providers for their households (Stites & Akabwai Reference Stites and Akabwai2010), but often do not control the means of production or profits beyond what is needed to cover subsistence needs (Caravani Reference Caravani2018). Simultaneously, inexpensive commercial grain alcohol has become more widely available, and many women report that domestic violence is most common when one or both parties are intoxicated. Evidence of the link between drunkenness and domestic violence comes from Karenga Sub-County in Kaabong. The Karenga Peace Committee banned the sale and consumption of grain alcohol, and rates of violence (both domestic and otherwise) reportedly plummeted.Footnote ⁴¹

A smaller number of sites in 2015 described a decrease in domestic violence.Footnote ⁴² Women reported that their increased mobility as a result of improved security allows them to leave home when their partners were drunk or combative. The ability to avoid the threat, regardless of time of day, means that women are able to de-escalate situations that might otherwise have resulted in violence.

Overall, our data show that high rates of domestic violence are the result of attitudes around gender relations at multiple levels. Male and female respondents describe that women are largely understood to be a man's property and are expected to obey their husbands. A woman is often blamed for situations beyond her control, such as children crying from hunger. These attitudes are reinforced by the local custom of publicly caning both husband and wife when the husband has been violent. Local leaders in one village in Kotido rounded up all the women in the middle of the night and threatened them with severe punishment should they continue to ‘refuse sex’ or speak back to their husbands.Footnote ⁴³