Resource management under climatic risk: a case study from Niger.

I. INTRODUCTION

Rainfall variation is often identified as the major risk faced by agro-pastoralists in the arid and semi-arid regions of sub-Saharan Africa [Swallow, 1994]. In these environments, households must adopt mechanisms to manage the variability in production of crops and livestock, and to mitigate the impacts of drought when it does occur. Among the many risk management strategies that have been identified, livestock mobility is often seen as one of the most valuable, since it enables herders to improve mean output as well as decrease output fluctuations associated with both spatial and temporal variability in rainfall [for example, Fleuret, 1986; Painter et al. 1994, Swallow, 1994; van den Brink et al. 1995]. Mobility is facilitated by the common-pool nature of most grazing resources, which significantly reduces the transactions costs associated with mobility [Niamir-Fuller, 1999]. But the common-pool nature of grazing resources also means that there are potential externalities, which lead to costs associated with resource management. These externalities, and the extent to which they are managed, affect decisions on stock densities observed on home pastures, herd mobility, and land allocation patterns. The purpose of this article is to develop an empirical model that incorporates the impact of both rainfall variability and costly cooperation on land use, land allocation and herd mobility decisions, and to apply the model to data collected in southwest Niger.

Issues surrounding the impact of climatic variability on the use and management of common resources, the vulnerability of rural households, incentives for privatisation, and conflicts among resource users are widespread throughout sub-Saharan Africa; and are in fact still quite important in North Africa, and West and Central Asia. In Niger, the government began implementing a new rural code in 1993 with the goal of re-defining the access, use, and management of natural resources in Niger [Secretariat Permanent 1993, 1997], though implementation remains stalled [Kirk and Ngaido, 2001; Ngaido, 1995; Gado, 1996]. While it is widely recognised that climatic variability is an important characteristic underlying the logic of the agro-pastoral system, designing a legal framework that addresses the need for flexible access while maintaining incentives to use and manage the resource has yet to be developed. Results in this article should help shed light on these wider issues.

A short review of the relevant literature and a discussion of the empirical context in Niger are presented in the next section. This is followed by the development of an empirical model of pastoral production systems in section III. In section IV, we discuss survey methodology and present descriptive statistics for sample communities. Results from model estimation are presented in the section V. We conclude by discussing policy implications and extensions to the existing model.

II. LITERATURE REVIEW AND BACKROUND ON NIGER

Recent literature on rangeland management in the arid to semi-arid regions has largely focused on the role of mobility in improving average production and reducing variability in that production. Van den Brink et al. [1995] present a theoretical model of livestock mobility and sedentary crop production that highlights the value of ex post adjustments to rainfall variability via herd mobility. There is also a large body of literature comprising the 'new range ecology' that also emphasises the value of mobility, based on the observation that forage productivity is overwhelmingly driven by rainfall [Sandford, 1982; Behnke and Scoones, 1993; Scoones, 1994]. Proponents further argue that, since stock densities have little impact on forage productivity, systems based on the accumulation of large herds and herd mobility to take full advantage of variations in forage productivity produced by variation in rainfall are highly efficient and equitable. A similar argument regarding accumulation is made in Livingstone [1991] and Fafchamps [1998]; with imperfect livestock markets, it may be preferable to build large herds in anticipation of a drought to ensure that more animals survive the drought.

It is interesting to note that none of the articles cited immediately above is concerned with the possible negative impacts stemming from overexploitation. The new range ecology literature largely argues that in these 'disequilibrium' environments, stock densities do not matter. Possible negative impacts of non-cooperation on the benefits of mobility or precautionary herd build-up at the community level is not considered in the models of van den Brink et al. [1995] and Fafchamps [1998]. In striking contrast, much of the previous literature on rangelands focused almost exclusively on the likelihood of over-exploitation and thus recommended privatisation of range resources; these analyses stemmed from Hardin's 1968 article on the tragedy of common pastures. Over-exploitation occurs when externalities are not fully internalised by group members. In these highly variable environments, there may be different types of externalities present, including crowding--or single period externalities directly impacting animal productivity, dynamic externalities that impact animal production indirectly through forage productivity, and additional externalities due to rainfall variability when producers are risk averse; we discuss these further in section III.