Village-level Behavior Under Conditions of Chronic Conflict

Abstract Drawing on a panel study of households established in 2002 and a revisit in 2008–2010 to a subsample, this paper explores the livelihood pathways of 24 households in 3 villages in Badakhshan in northeastern Afghanistan. It finds that most households were worse off than they were in 2001, although they experienced a brief period of relative prosperity based on the 1 market choice available, opium poppy. The paper draws attention to the corporate nature of villages and their variable capacity to support the provision of village-level public goods. This variability is influenced in part by the relative richness of the resource base of the village and the related degree of social differentiation. Where land inequalities are high and the elite are economically secure, they have few incentives to widen provision of public goods and can be immune from social sanctions. Where the elite are economically insecure, they are likely to have a shared interest in supporting village solidarity and a moral economy and may promote the provision of public goods. External interventions focusing on village governance need to pay much greater attention to village preconditions given the extent to which the effects of such interventions are often subject to the behavior of the elite and preexisting customary structures.


Introduction
This paper is an inquiry into household livelihood trajectories (Bagchi et al 1998;Murray 2002) under conditions of chronic conflict and the influence on these of village context and its social order (North et al 2007: 47). Although there have been a multitude of village-level studies in South Asia, including Afghanistan (Tapper 1991;Coburn 2011), few have established panel data sets and tracked households over time or situated trajectories within a wider institutional context (Da Corta 2010). Even fewer have compared villages, and little attention has been paid to intervillage variation (Wade 1988 is a notable exception). As Srinivasan (2004: 78) noted in his analysis of long-term change in villages in southern India, ''higherorder statements about the developmental experience of villages is notable for its absence,'' although such differences are widely recognized.
A village is a place, but as with markets-which are places and also institutions with rules that govern behavior and the conditions and terms under which exchanges take place (Harriss-White 2003)-the village contains a multitude of institutions that govern the behavior of people inside it. Where the state is largely absent, as in Afghanistan, such institutions continue to play a key role.
Three customary structures are commonly found in most Afghan villages (Noelle-Karimi 2006;Brick 2008) that are central to village governance and the provision of basic public goods, including those of security and dispute resolution (Smith 2009). These are the village council or shura, the village leaders (maliks or arbobs), and the village clergy (mullahs). Each of these customary structures has distinct and nonoverlapping areas of authority (Brick 2008), gaining authority and legitimacy from different sources. In the case of the village council, its membership is based on reputation and performance earned through managing dispute resolution. The village headman is the key interlocutor between the village and district authorities, usually selected by the village. The clergy's authority is derived from religion and speaks on matters determined by Sharia law. As will be argued, the 3 study villages are distinctly different in the performance of their customary structures, with-it is suggested-long-term effects for the welfare of their households.
The paper first describes the methods and context of the study before exploring livelihood trajectories and village differences and their effects. The final section develops an explanation of what underlies variability in village behavior, drawing on the contrasts between them.

Systems knowledge
Mountain Research and Development (MRD) An international, peer-reviewed open access journal published by the International Mountain Society (IMS) www.mrd-journal.org

Methods
In 2002 a panel set of 390 households in 21 villages was established in 7 contrasting provinces to build understanding of Afghan rural livelihoods post-2001 (Grace and Pain 2004). A subsample of these (64 households in 8 villages in 3 geographically contrasted provinces) was revisited in a second round restudy in 2008-2010 to explore household livelihood trajectories and welfare outcomes in the intervening period (Kantor and Pain 2011: 9-13). This paper reports on the livelihood trajectories of 24 of those restudy households living in 3 villages in Badakhshan, linking these with a detailed examination of the factors underlying village variability. The 3 study villages were originally selected in 2002 for their differences in resource endowments, but, as became clear in the study, they also had major differences in the ways in which customary institutions functioned.
The restudy focused on investigating patterns of similarity and difference in household trajectories and used a qualitative approach to collect in-depth information from a small number of carefully selected contrasting household case studies (Flyrbjerg 2006), based on a wealth group classification done in 2002. This provides a depth of understanding that statistically representative approaches cannot achieve (Hall 2003).
The field teams applied retrospective in-depth interview techniques to explore household lives and livelihoods from 2002 to 2009 to understand decision-making, resulting strategies, and their impacts in response to changes in circumstances. Interview teams were composed of 2 female and 2 male Afghan interviewers, allowing men and women of the respondent households to be interviewed separately. Two interviews were held with men and 2 with women, giving a total of 4 interviews per household. Further interviews were then held by the authors with key informants in the villages and districts to explore village social structures and differences between villages.

Badakhshan
Located in the northeast of Afghanistan, the mountainous province of Badakhshan is remote, economically marginal, and grain deficit, and there has been a history of seasonal and long-term migration. The provincial poverty rates (greater than 58%) remain among the highest in the country (MoE and WB 2010: 28).
There was an early investment in education from the 1950s so that education levels in the province in 1978 were notably higher than in other provinces (Guistozzi 2010), leading to the emergence of an educated elite. After 1978 there was considerable conflict between the leftist government and the mujahideen, and all the study villages experienced fighting during this period, although to different degrees. With the capture of Kabul by the mujahideen forces in 1992, Professor Burhanuddin Rabbani and Ahmad Shah Massoud-the 2 major northern political players-occupied key positions in the government. When the Taliban captured Kabul in 1996, Rabbani and Massoud established themselves in Badakhshan and the Pansher, a historical point of access to Badakhshan, as the center of opposition to the Taliban, holding out until the Taliban lost power.
Up to 2005 Badakhshan experienced considerable anarchy (Guistozzi and Orsini 2009) under the weak patrimonial rule of Rabbani, with armed parties jostling for power and control of the rising opium economy (Goodhand 2000;Pain 2008). From 2005 President Karzai increasingly intervened in the province, seeking to ''replace local systems of power and patronage…with an alternative one dependent on Kabul'' (Guistozzi and Orsini 2009: 41). He supported a local politician, Zalmay Khan, who has had a long history of shifting political allegiances to build his political and economic power base, and the province has remained subject to political conflict, reinforcing insecurity.

Three villages: Toghloq, Shur Gul, and Khilar
The 3 study villages (Table 1) are located between 1200 and 2000 masl ( Figure 1). Toghloq, the lowest-altitude and most accessible village, is 1 of 5 villages in a well-irrigated lateral valley that feeds the main Jurm valley. It has the highest proportion of irrigated land of the 3 villages and the greatest land ownership inequalities. Earlier, most households found employment as farm laborers, but there is now significant outmigration for employment, including joining the army or police.
Shur Gul is the largest and remotest of the 3 villages, being 3 hours' drive from Jurm in a narrow plain in the Kokcha valley. It has always been grain deficit, is unreliably irrigated, and has a long history of labor migration. In the past, the neighboring lapis lazuli mines provided employment for many households in the village, but this is now no longer the case. This village has had a long history of education. Khilar, the smallest of the 3 villages, is located on a small plateau situated above a narrow valley. It is inhabited by a religious minority and with limited land resources; historically it has been economically and politically marginal. There has been a long history of male migration for work.

The household trajectories
The economic life of the study villages has been subject to considerable vicissitudes. War brought destruction and created refugees, although the extent of both was less in Badakhshan than elsewhere. Drought from the late 1990s affected the mainly rainfed economy of Badakhshan, bringing food insecurity and distress migration. An opium economy that gathered pace from 2001 onwards brought a level of prosperity to the province beyond its historical experience (Pain 2008), relieving debt and enhancing food security. The decline of this economy from 2006 onwards (although it has shown signs of a return; see Pain 2011; Figure 2), driven by falling prices and counternarcotic pressures, led to a sharp drop in the health of the rural economy compounded by recurrent droughts and volatile food prices.
All case households benefitted from the opium economy during the period of 2000-2006, but Toghloq benefitted most because of its resource richness. These were the years of relative prosperity, with high levels of food security for most. Since then there has been a decline in rural employment and wage rates, poor rainfall, and a significant rise in grain prices during 2008. Drawing on what households reported about their change in circumstances, they were classified as ''prospering economically,'' ''coping,'' or ''declining'' (Pain 2010: 27). Out of the 24 households, 3 had prospered (Table 2) since 2002. Another 3 had managed to cope, broadly maintaining their economic position. The majority (18) had seen a decline in their economy, indicative of the precariousness of life in the province. Of the 3 prospering extended households, 1 in Toghloq had done so through agriculture, but was originally landless. Rich in male labor devoted exclusively to sharecropping land, the opium economy allowed the household to sharecrop on additional land at the peak of opium's profitability and leverage that wealth into land purchase. The 2 households from Shur Gul have prospered primarily through salaried employment, 1 as a teacher and the other driving for a nongovernmental organization (NGO).
The 3 coping households-2 from Toghloq and 1 from Shur Gul-have either benefited from inherited resources (land) or through education secured some form of salaried employment. But large family sizes have raised consumption demand close to the capacity of these households to meet that demand.
For the declining households, variable but combined constraints of land, labor, household-specific health events (sickness and death), and aging contributed to their economic decline. In 1 of the Shur Gul wealth group II households, through a combination of bad luck and internal conflict that led to the son's setting up a separate household, there has been a severe economic decline (Pain 2010: 28). Khilar, which lost valuable pastureland to powerful commanders during the war, has suffered a major decline in its livestock population, and this has continued since the ending of the opium poppy cultivation. Many households have moved into food rationing as grain prices rose in 2008 and diversified into low-return and uncertain activities such as brushwood collection. Although some households have survived on informal credit, the level of poverty is such that in Khilar poorer men have been unable to marry. The sending of sons to join the army or police has its risks, and 1 household from Khilar had a son killed after he joined the national army.
The overall picture of economic life over the last decade for the study households is one of decline, thus limiting the contribution of individual action and markets to household welfare. Although there are resource differences between villages, more significant to the longterm prospects of households are the differences in village behavior and the extent to which collective action has provided public goods. Thus although the short-term prospects for the Shur Gul households are limited, in the longer term these households are better placed than the poor from the other 2 villages, given the investments the village has made in education and other public goods.

Village social orders
In the disputed Afghanistan presidential election in 2009, when President Karzai gained a second term, his representatives and those of his chief opponent Abdullah Abdullah came to Badakhshan and visited each of the 3 villages. Toghloq largely ignored the election, had little to do with both campaign teams, and the campaign teams in turn did not pay much attention to the village. Shur Gul, by contrast, was visited by representatives of both campaigns, who each asked to set up an election office in the village. Both requests were considered by the village council, and both were rejected on the grounds that the presence of either party might contribute to conflict in the village during and after the election. In Khilar, President Karzai's representatives took the village leaders and local power holders to campaign on the president's behalf in the neighboring valleys. The above account was given by 1 key informant and corroborated in the 3 villages. Drawing on evidence from field observations, key informants, and village-level discussions (Pain 2010: 16-24), it appears that the actions of the leadership of Shur Gul to manage the village's external relations in the interests of the village were consistent with other actions they had taken in the past.
The ways in which the other 2 villages engaged in the elections were also in keeping with their past behavior. This observation on the contrasting ''behavior'' of 3 villages begs the questions of what is meant by village behavior, what underlies it, and what effects it has on the security and welfare of people who live in these villages.
Central to an understanding of the social order of the 3 study villages is the behavior of their elites. The villages differ significantly in this respect, with implications for household access to public goods. The first point of contrast is with respect to education (Table 3). In the case of Shur Gul, over 60 years ago the arbob (village leader) put the village on an educational track through the establishment of a school. A significant number of boys graduated and went to university in Kabul and elsewhere, building a wider network of social connections. After 1978 many returned to the village to maintain its educational investment and were paid for by the village. A girls' school was established in the mid-1990s and the first girls were graduating by 2009. Toghloq also established a school, but there was less interest from the elite in education, and by 1978 only a few students graduated to university. After 1978, the village and the valley as a whole quickly joined the opposition. The school, as a symbol of government presence, was destroyed by the mujahideen, the teachers were killed, and all education was stopped until 2001. The school was restarted after 2001, but recruitment of both boys and girls has been gradual. Khilar has never had its own school, and boys went to school in the valley. A few managed to graduate before 1978, but access for girls to education did not start until after 2001, and the distance still restricts girls' access when they are older because of social norms about such girls being in public. Table 3 compares-to the advantage of Shur Gul-the educational outcomes between 3 villages. In contrast to the other 2 villages, all the poor households in the Shur Gul sample were making every effort to keep all their children in school because of the long-term advantages they saw in education.
The differences in educational outcomes in the 3 villages are reflected in the levels of other public goods. The village leadership in Shur Gul has been effective in securing levels of health provision, safe drinking water, road access, and electricity that the other 2 villages have not reached (Pain 2010), in part because of its effectiveness in building connections to NGOs, as discussed below. Khilar, in contrast, only got road access in 2007.
The second point of comparison is with respect to security. Toghloq entered into armed opposition and the valley successfully defended itself against government forces after 1978. With the departure of the Soviets the valley came under control of a powerful regional commander, allowing various subcommanders within the valley to compete for position. Collective action had been effective in defending the village and valley from the outside world during conflict. Even in 2005, when an opium eradication team attempted to enforce the opium ban, it was met with armed resistance and the team's vehicles were seized and burnt, an action that would have been unlikely in the other 2 villages. But the power of commanders within the valley has been a source of insecurity, and 1 informant reported how his daughter was taken by force to marry a commander's son (Pain 2010: 43).
Khilar suffered predation and lost pastureland to powerful commanders in the valley during the war (Pain 2010: 21). Only when the village placed itself under the protection of a sympathetic commander in the valley did they gain protection. This valley commander still retains influence over the village, and his authority exceeds that of the village council (Pain 2010: 23).
In contrast, Shur Gul at the start of the conflict selected an educated and prominent figure to lead the village defense and managed to limit its engagement in the conflict. Although the village had divisions, its ability to maintain good external relations as a form of defense was central to its survival. Links with key provincial commanders established by the educated elite ensured both its physical and economic survival, particularly during the drought when they gained access to work in the lapis lazuli mines for a regular period each year. The ability of the village leadership to build external relations was critical to gaining the interest of NGOs after 2001 and the expansion of public good provision, with long-term welfare effects for its inhabitants in contrast to the other 2 villages.

Discussion
What underlies the contrasts in behavior of the leadership of these 3 villages as seen by their actions in the recent presidential election and their different interests in supporting the public good in their villages? Brick (2008) argues that 3 features of village customary structures in addition to their distinct areas of authority can be supportive of village-level public good provision. The first is the capacity of these customary structures to independently raise revenue from within the village. The ability of the shura in Shur Gul to raise a levy for the school teachers during the war is consistent with this. Second, there need to be checks and balances that can prevent abuse of authority by anyone of these customary structures. The reason for this relates to the third factorthe need for there to be sufficient actors who have the ability to stop potential abuses of power and act as veto players. Where land distribution is relatively equal between landowners, power is more likely to be dispersed. Where landownership is concentrated, so is power. Under such conditions there are few constraints on elite behavior and they are more likely to act in their own interests.
It is the relation between social solidarity and inequality that is the critical issue. Where land inequalities are higher and the elite are more economically secure, as in Toghloq, they have fewer incentives to widen provision of public goods and are largely immune from social sanctions; social relations can be hierarchical and exploitative. Where, however, the elite are economically insecure, as in Shur Gul, they are likely to have a shared interest in supporting social solidarity and a moral economy (Scott 1976) and promoting the provision of public goods. But Khilar, with its small size, limited resources, and social marginality, indicates the limits to village collective action.
Two conclusions may be drawn from this discussion. First, the evidence suggests that collective action is likely to be stronger under conditions of relative equality linked to poorer resource conditions. Second, village preconditions in Afghanistan are likely to matter to program design. Although programmatic practice has been attentive to the need for provision of public goods at the village level, it has specifically sought to displace existing village customary structures (Pain and Kantor 2010). There has been no interest in understanding and responding to the variability in village preconditions that might affect public good provision. There has also been no attention to the ways in which ''modernizing'' organizational practices have engaged with customary institutional behaviors. The consequence of this is that interventions have more often operated subject to existing practices rather than displacing them, a process of institutional ''bricolage'' (Jones 2009). Sometimes this has worked to the good, expanding public good provision, as in Shur Gul. In others it has not, reinforcing elite positions and leading to external resources being captured by the elite.