Analysing the context: actors, dynamics and trends in violence

1Why analyse the context

All security decisions rest on an understanding of the context. A weak analysis produces two symmetrical errors: underestimating a threat (and exposing staff) or overestimating it (and abandoning populations who could have been helped).

A good analysis answers a few questions: who controls the territory and the roads? What are the actors' interests, and how does humanitarian action affect them? What types of violence exist, against whom, where, when and why? What is changing?

Sources: national staff (the first source, often the most nuanced), communities, authorities, other organisations, NGO security coordination forums, the United Nations (UNDSS), local media, social media (with caution), public conflict databases. All information is cross-checked and rated according to the reliability of the source and the credibility of the information.

2Analytical frameworks

Two simple frameworks structure the analysis:

• Adapted PESTEL: political, economic, social, technological, environmental, legal, to which security is added (armed groups, crime, security forces).
• Historical analysis: a timeline of major events over recent years, to identify cycles (fighting seasons, election periods, harvests, transhumance).

The analysis leads to scenarios: the most likely, the most dangerous, and possibly an improvement scenario, each with observable indicators that signal it is approaching.

Example of indicators. Scenario "resumption of fighting around the town": families moving south, closure of weekly markets, sudden rise in fuel prices, departure of foreign traders, reported troop movements, hostile rhetoric on local radio stations. Three out of six indicators observed trigger a crisis meeting and a review of the security level.

3Mapping actors

Actor mapping lists those who can influence security and access: national and local authorities, armed forces and police, non-state armed groups, traditional and religious leaders, militias, criminal networks, traders, media, host and displaced communities, other organisations.

For each: influence (low to high), attitude towards the organisation (hostile, neutral, favourable), interests and current relationship.

A hostile actor with high influence is a priority for the analysis and for the acceptance strategy, not just a threat to be avoided.

4Reading incident data

Incident data help to objectify the threat, provided they are read correctly:

• Raw frequency: number of incidents per period.
• Rate: incidents relative to exposure (number of staff, trips, person-days). Without a denominator, an increase may simply reflect an increase in activity.
• Reporting bias: an increase may come from better reporting, not from a deterioration.
• Typology: the same total can hide a fall in thefts and a rise in kidnappings, which are far more serious.

Example. Year 1: 24 incidents for 300 staff. Year 2: 36 incidents for 600 staff. Rates: 8 versus 6 incidents per 100 staff per year. Despite a raw increase of 50%, the rate has fallen by 25%. But if serious incidents (violence against people) go from 3 to 9, their rate goes from 1 to 1.5 per 100 staff: severity is increasing. The reading combines rate and severity.

5Understanding violence against aid workers

Attacks on aid workers follow several logics, which must be distinguished because they call for different responses:

• Criminal: financial gain (thefts, robberies, kidnappings for ransom). Protection measures and discretion about valuables are effective.
• Political or military: the organisation is perceived as an ally of one side, or its action disrupts a strategy. Acceptance, perceived neutrality and communication are central.
• Linked to the work itself: targeting disputes, contested recruitment, dismissals, unaddressed complaints. Programme quality and accountability are security measures.
• Incidental: being in the wrong place at the wrong time (crossfire, shelling, mines). Analysis of areas and timings, movement procedures.

Many serious incidents affecting NGOs are linked to internal or neighbourhood tensions (dismissed staff member, rejected supplier, excluded beneficiary): security analysis must also look at the organisation itself.

6Producing a useful analysis

A useful analysis fits on two pages, is updated regularly (monthly in an unstable context, after any major event), and ends with concrete implications:

• areas and roads that are accessible, accessible under conditions, or off-limits;
• travel times and modes of transport;
• actors to meet and key messages;
• indicators to monitor and review thresholds;
• consequences for programmes (timetable, modalities, remote management).

It is shared with the people who need it, while respecting confidentiality: an analysis that names armed actors or informants can put people in danger if it circulates. Nor should it use confidential information from other organisations without their explicit agreement.