I’m researching the impacts of age structure on retail sales and other measures of economic activity at subnational levels. In many cases I find states are too large an area to look at for meaningful insights. However, I hoped for more from a graph than this.
That is a pretty big cluster there and there really is not much to do with that. No clear relationship at all. Even the outliers are not that exciting. So scratch dependency ratio as far as a meaningful impact on retail sales. Time to move on to other age-related population measures.
On another note that dependency clustering is pretty interesting in its own right. At some point I will need to see if the counties outside the cluster have some geographic clustering or other aspects.