Cointegration testing may lead to false correlations with serious societal implications.
Cointegration testing can't always tell if a relationship between two things is real or just a coincidence. Even when two things seem to be connected, like NASA and suicide rates, or margarine and divorce rates, it might just be a fluke. The study shows that cointegration testing can sometimes miss real connections and mistakenly find fake ones. So, it's important to use common sense and other information along with cointegration testing to figure out if a relationship is meaningful or not.