The Enemy Within: The Impact of Corruption on Security Force Assistance

International Security Assistance Force commander General David McKiernan visits national policemen in Afghanistan in 2009. Courtesy of Lance Cpl Brian Jones / Wikimedia Commons

With corruption playing a significant role in the outcome of recent security force assistance missions, a different approach to tackling the problem is needed.

An issue that haunts more or less all security force assistance missions to various degrees is widespread and large-scale corruption in host institutions. Such corruption can be serious enough to doom entire security apparatuses to dysfunctionality. The recent exposure of macroscopic cases of corruption in the Afghan security forces follows the 2014 exposure of another macroscopic case within the Iraqi security forces. The collapse of the Iraqi army in the face of the 2014 offensive of the nominally much smaller forces of the Islamic State was a big shock, even if it does not appear to have led to a serious rethinking of security force assistance in Afghanistan. Seven years later the Afghan security forces suffered an even worse collapse, from which they did not recover.

The fact that corruption affecting smaller security force assistance missions has not made the headlines does not mean that it does not occur there too. The extent of corruption in Mali’s security forces, for example, was the topic of a 2015 report by Transparency International. The impact of corruption on security force assistance in Somalia is also well known.

Back in 2019, Jahara Matisek and William Reno quoted a Senegalese general who, commenting on why the US has been struggling to help establish effective militaries around the world, pointed the finger at the local political economy: ‘The logic of their politics will show you the quality of their military’. Matisek and Reno concluded their analysis by arguing that ‘The basic dilemma for providing SFA to weak governments is that the beneficiaries are often implicated in the activities that this assistance is meant to address’. The expectation that corrupt actors will reform themselves certainly seems unrealistic. So, is reform impossible, short of taking over and establishing a protectorate?

In a context of weak or illegitimate institutions and a fluid or missing political settlement, the very concept of corruption risks being an artificial superimposition by external actors. Entering such operational environments with the expectation that corruption will be a marginal factor risks being as much of a delusion as entering a counterinsurgency operation with the expectation that the enemy will play by the ‘rules of war’.

Without a solid political settlement, the right of a political elite to extract rent in the guise of taxes will always be contested, as will the use of those taxes. Similarly, ‘corruption’ is often just a label that rival elite factions attribute to each other. What external observers may interpret as ‘corruption’ are in fact power struggles over resources. If power is arbitrary, so is tax collection. As Iraq and Afghanistan demonstrated, the endorsement of industrial democracies is not enough to offset a fluid political settlement.

A capable government with solid institutions, a functioning legal system and the ability to effectively centralise revenue extraction does not usually need security force assistance from abroad. Missions to places that lack all or most of these are likely to remain the norm. These missions should adopt the working assumption that they will face extensive ‘corruption’ (or disputed revenue extraction) affecting the security sector, and that this ‘corruption’ will only get worse if extra resources are pumped in from abroad. This is why, as noted by Astri Surkhe some years back, ‘more’ can indeed be ‘less’.

To accept corruption in the environments surrounding security force assistance missions does not mean that mitigating measures should not be taken

To some extent, corruption has to be accepted as a fact of life by security force assistance missions, at least until security sector and political reforms start yielding such dividends that ‘corruption’ effectively becomes corruption without inverted commas and can be tackled. Such reforms unfortunately do not occur successfully very often.

To accept corruption in the environments surrounding security force assistance missions does not mean that mitigating measures should not be taken. Corruption tends to be encouraged when those handling the money (typically politicians and bureaucracies) and those having to deliver the results (in these contexts, those fighting on the host government’s side) are far apart. The consequences of embezzling funds are not shared by the beneficiaries, at least not immediately. The presence of a security force assistance mission might even reinforce the feeling that corrupt actors will be able to get away with their behaviour without suffering consequences such as regime collapse, let alone soldiers and police officers getting slaughtered in the provinces.

The more centralised and bureaucratised a security force is, the larger the scope for corruption, as long as the interests of the bureaucracy and those of the combatants diverge. Could the segmentation of security forces into smaller entities represent an antidote to corruption, simply by reducing centralisation and bureaucratisation?

We can draw a parallel here between the analysis of Mushtaq Khan and others on how patrimonialism can be developmental under certain conditions, and the processes that might contain corruption in security force establishments. In developmental patrimonialism, long-horizon rent management translates into incentives for behaviour leading to developmental outcomes. The point of segmenting security forces into smaller entities is really about changing the incentives of the key actors, creating long-horizon attitudes. As long as those running the system have enough at stake in the success or at least the survival of those fighting it out on the battlefield, their incentives to participate in large-scale corruption of the type that makes a system dysfunctional will be reduced.

Strong institutions are essential for running large organisations successfully. In their absence, the incompatibility of patrimonial ownership and large bureaucratic organisations turns the latter into no man’s lands, prey to factionalised elites who are desperate to find the resources for pursuing their power struggles. The result is large-scale corruption.

Smaller organisations, while in principle lacking the economies of scale of larger structures, are easier to control patrimonially. From the perspective of a security force assistance mission sent from some advanced economy, this would normally be seen as a negative, but this is where developmental patrimonialism kicks in. If members of the national elite can assert their ownership over a security establishment, they are more likely to take care of it.

As long as those running the system have enough at stake in the success of those fighting it out on the battlefield, their incentives to participate in large-scale corruption will be reduced

Even when owning security establishments patrimonially, actors might still want to divest resources for use elsewhere, or for personal enrichment. This could well be the case if their ‘ownership’ is unlimited. And in such a case, even if they decide that they need to keep their institutions functional and efficient, and therefore have every interest in preventing corruption, actors may wish to turn them into personal fiefdoms, serving narrowly defined interests and contributing little or nothing towards ending the conflict.

The problem becomes one of either preventing ‘complete ownership’ of security establishments by local elites, or manipulating the incentives of the ‘owners’ of these smaller security establishments, pushing them to identify their own interests with the interests of the wider security system and of the state.

Is this something that a foreign security force assistance mission could do successfully? Preventing elites from seizing complete control of security establishments might be a conditionality that relatively large and well-resourced security force assistance missions could impose, if they could recognise the problem early on. Deploying incentives has certainly been done before, and sometimes on a massive scale (see Afghanistan). But the Afghan case also suggests that skilfully and successfully deploying such incentives is not something that even the largest missions can do easily.

What is needed is to give up on trying to establish or sponsor large bureaucratic organisations where the political environment is not conducive and human resources are scarce, and to instead aim to develop and coordinate smaller organisations towards a common end. The set of skills required may well differ from those typically associated with security force assistance missions so far, and will certainly require much greater flexibility, but for such missions to be more successful in the future, some change and experimentation is in order.

The views expressed in this Commentary are the author’s, and do not represent those of RUSI or any other institution.

Have an idea for a Commentary you’d like to write for us? Send a short pitch to and we’ll get back to you if it fits into our research interests. Full guidelines for contributors can be found here.


Dr Antonio Giustozzi

Senior Research Fellow

Terrorism and Conflict

View profile


Explore our related content