Why COVID-19 is impacting security environments differently

By: Tom Mather

According to data collected by Standard Risk, 79% of security incidents related to COVID-19 in East and Central Africa, between 1 March - 7 May, took place in Kenya, Malawi and Uganda. This blog piece seeks to explain why is this the case and suggests that countries with higher levels of democracy may face more civil unrest following the enforcement of COVID-19 restrictions.

In monitoring COVID-19-linked security incidents across 13 countries in East and Central Africa, we see that 79% of the total number of such security incidents have occurred in Kenya, Malawi and Uganda. The remaining 10 countries have each witnessed 5 or fewer COVID-19-related security incidents. Most commonly, incidents are either protests against restrictions or actions taken by the security forces to enforce curfews and new trading rules. We have not seen significant harassment of expatriates or aid workers, nor have we seen evidence that COVID-19 is impacting patterns of armed conflict within the region.

To understand this observation, we took two variables that could be expected to have a direct impact on the prevalence of COVID-19-related security incidents: 1) the level of restrictions imposed by governments in order to halt the spread of the virus, and 2) the level of freedom in each country. These were chosen as the imposition of restrictions changes the operational environment, and the level of freedom can be used to indicate how well a population will adhere to new regulations.

Taking data from the stringency index of the Oxford COVID-19 Government Response Tracker (OxCGRT), combined with the number of COVID-19-linked security incidents recorded by Standard Risk, and overlaid with ratings from the Economist Intelligence Unit (EIU)’s Democracy Index 2019, we get the below chart:

X-Axis: COVID-19-linked security incidents (1 March-7 May 2020); Y-Axis: OxCGRT’s Stringency Index; Colour palette: EIU Democracy Index 2019.

Observations

Levels of stringency – or restrictions – do not appear to directly correlate with increased violence. Eritrea and Djibouti have comparable levels of restrictions to Kenya and Uganda, and yet have far fewer incidents. When you factor in levels of freedom, however, a correlation can be observed. The three countries with the most incidents also score high on the EIU’s Democracy Index. Although correlation does not mean causality, the chart suggests that government-imposed restrictions can lead to increased security incidents, but only where a certain level of freedom exists. To explain this outcome, those in authoritarian states are presumably less willing to protest or violate restrictions as they are aware that the security forces will be effective in any kind of response. In non-authoritarian states, there may be a belief that restrictions are more ‘optional’, and that restrictions can be challenged.

Digging into the types and dates of incidents recorded, several patterns are also observable. The vast majority of incidents in Malawi (83%) are linked to protests ahead of the expected imposition of a 21-day lockdown starting 18 April (the government’s decision was subsequently blocked by the High Court). Meanwhile, in Kenya and Uganda, most security incidents (protests and curfew/trading enforcement by the security forces) took place in the first two weeks after restrictions were imposed. It is perhaps not surprising that the evidence suggests that covid-19-linked security incidents are highest immediately before and after the imposition (or planned imposition) of restrictions. Restrictions represent shocks to the status quo, which heighten the likelihood of security incidents. Over time, however, as restrictions become the new normal and a new equilibrium is established, the likelihood of security incidents drops. This has been seen in both Kenya and Uganda where incidents peaked immediately after restrictions were imposed, then faded (although have not disappeared entirely).

Why does this matter?

If this assessment holds true, then we can argue that:

• Changes in government restrictions in non-authoritarian regimes tend to prompt an increase in security incidents. This works for the East and Central Africa region, and it may work elsewhere.
• In the coming weeks and months, three countries should be watched closely. Malawi could see a spike in incidents if the government re-attempts to impose or succeeds in imposing a lockdown. If Tanzania were to impose significant restrictions—which given the current trajectory seems improbable—it is likely that incidents would increase given the level of freedom in the country. With a recent history of heavy protests, Ethiopia deserves special attention. Although classified as ‘not free’, it is closer to being ‘partly free’ than other countries in the category. Presuming the pandemic continues to negatively impact local economies, Ethiopia may be particularly vulnerable to social unrest over time.
• For Ethiopia, Malawi and Tanzania, the announcement and imposition of restrictions can be seen as destabilising factors, and accordingly could be considered as triggers for contingency planning. By contrast, increased restrictions in Burundi could be less likely to provoke a rise in security incidents, making increased restrictions less applicable as a trigger for contingency planning.

Caveats

The assessment is based off a small data set, using results from 13 countries. Moreover, a multitude of factors influence the likelihood of the occurrence of security incidents, including precedents, norms and cultural factors. Over time, other factors may alter the results of this assessment. Economic hardship, for instance, is an all-but-inevitable consequence of the pandemic, and this will likely further alter the existing security environment. There is also a considerable feedback loop where heavy-handed government response can lead to protests, which in turn can lead to a further government response. Additional research is required to determine whether the same observations and conclusions can be applied outside of the East and Central African region.

For additional information on the data used in this analysis, please see the Document Version.

About the author

Tom Mather is a former humanitarian security manager, having worked in regional safety and security roles for NGOs in the Middle East, Afghanistan-Pakistan, and East Africa. He is the technical director at Standard Risk, a security analysis company registered in the UK and Kenya. Standard Risk’s focus is supporting organisations to make better security decisions by providing clear and actionable security information and analysis. The team is comprised of analysts, ex-humanitarian workers and security professionals, who are passionate about using data and analysis for better, safer decisions.

Related: