I think it's why inequality plays a role and not just poverty itself.
Extreme poverty puts people in desperate situations, but I feel (this is just a supposition, I've not done nor read research) that people can also recognize circumstances. I've less impetuous against my neighbours if they are struggling in similar ways than I may be. There may also be other mitigating circumstances such as whether or not people empathize with each other's positions and provide any sort of community support, even if that support is low in absolute wealth.
Doing a search I found this (bold emphasis mine): https://www.worldbank.org/en/news/press-release/2021/10/07/south-africa-a-new-social-assistance-assessment-aims-to-help-strengthen-policies-and-programs-for-the-poor
"The report finds that overall, South Africa’s system of programs is effective, well-targeted, and provides sizeable benefits to the poorest households. The social assistance system effectively reduces poverty and inequality rates. Cash transfer programs are having positive and important impacts on a wide variety of outcomes, including nutrition and food security, educational attainment, health, labour supply, and livelihoods. Furthermore, the grant system has established effective delivery systems that identify beneficiaries and include registration, payment, and grievance redressal processes which can be scaled up rapidly during crises such as the COVID pandemic."
What's interesting is that this article talks about how social assistance reduces both poverty and inequality rates. I'm not sure the intricacies of how Gini coefficients are calculated, but it would be very interesting if the world leading inequality rates assessed by the earlier link still took into account the social services. That is, are the inequality rates still the highest in the world even with some of the more generous social services in Africa? Because that could mean the circumstances within the country would be even worse without any of the social safety net that the World Bank says is helpful.
EDIT: Also interesting is there are gaps in data for crime rates in a lot of the poorer countries in Africa.
When looking at Macrotrends.net many of them had their most recent data put in 8+ years ago. It's possible that the poverty levels of the country make assessing the criminality almost impossible?
For reference, same link is able to show poverty rates for many of the same countries.