If you make composites where all the parts are intercorrelated and the weights aren't degenerate, even if the items are still disjoint, they quickly come to correlate extremely highly with one another. This is part of why GDP per capita correlates so highly with everything good.

In fact, this is part of why many composite variables like test scores, health quality indices, consumption measures, and so on tend to be so highly correlated: It's just a mathematical quirk of aggregation when you make composites! Compare that to ecological aggregation.