International policy-makers, journalists and researchers often use the term »warlordism « to describe the physical insecurity in Afghanistan. The perception dominant here is that warlords are a counterweight to the state and that the sole motivation of warlords lies in their personal economic profit and enrichment. However, the variety of actors and security arrangements found on the local level in Afghanistan reveals that this general understanding of warlordism is not sufficient on its own. This is explained with reference to the examples of the provinces of Kandahar, Kunduz and Paktia.

In Kandahar province the combination of insurgency, well-financed drug networks and hierarchical tribal structures have in the course of recent years restricted the influence of the Afghan government and instead favored the emergence of very few, strong warlords as well as the return of the Taliban movement. By contrast, the province of Kunduz harbors a high number of warlords, who differ widely in their scope of influence. Moreover, what was observed here is a complete lack of religious, ethnic-tribal or even modern institutions capable of constraining the arbitrary acts of the rulers. This can be ascribed to the country’s changeful history, and it results in a strong localisation of the »rules of the game« and varying architectures of the power structure. In Paktia, the »pashtunwali« – the Pashtun code of honor and law – has sustained an overall institution which rules the daily life of all tribesmen. Thus the Pashtun egalitarian ideal constrains the behavior of tribesmen, and tribal elites are controlled by checks and balances. It is for this reason that the tribal system in Paktia has restricted the emergence of warlordism.

These examples make it evident that the term »warlord« is not sufficient to describe the current (in)security in Afghanistan. Neither is there a uniform type of actor nor can these actors be accurately described by a single term. However, alongside the variety of actors, regional differences play a crucial role. Moreover, it would be misguided to position these security architectures on a linear spectrum between the poles of the state on the one side and the warlords on the other. Interestingly enough, warlordism arose especially in those regions in which the state as such is perceived as an important point of reference and as a political resource, while in the tribal regions warlordism has remained more or less absent. This is consistent with Charles Tilly’s argument that the emergence of warlordism is a constitutive element of state-building processes.