Fig. 3: Regional-level inequality outcome.
From: Local Network Interaction as a Mechanism for Wealth Inequality
The figure displays the regional-level inequality outcome by izigodi (n = 23). We first compute a value indicating the percentage increase in regional-level v inequality score over time (\(\Delta {{Inq}}_{v}={{Inq}}_{v,2018}-{{Inq}}_{v,2016}\)). We then plot the distribution of this value against the regional-level inequality score at baseline (\({{Inq}}_{v,2016}\)), categorising regions into those with a stable/decreased inequality score (\(\Delta {{Inq}}_{v}\le 0\)), those with an increase ranging from about 0-2% (\({0 \, < \,\Delta {Inq}}_{v} \, < \, .02\)), or those with an increase greater than 2% (\(\Delta {{Inq}}_{v}\ge .02\)). The value is then binarized as our regional-level outcome variable that indicates whether the area had experienced an increase in its inequality over time (\(\Delta {{Inq}}_{v}\, > \,0\)). To ensure the increase is not driven by other demographic changes in this highly mobile population (e.g., population size), we restrict our analyses to regions (n = 6) with an increase in inequality score distributed at the upper 75th percentile (\(\mu+(0.675)\sigma\)), equivalent to more than a 2% increase (\(\Delta {{Inq}}_{v}\ge .02\)).
