Thursday, October 14, 2021

Here We Go Again

Long-time readers know a hobbyhorse of mine is the misuse of correlational studies to claim causation. I have repeatedly harped that if A and B are correlated, all we know is that they tend to occur together. We do not know whether A causes B, B causes A, or some unspecified variable (or combination of variables) C causes both A and B. 

RealClearInvestigations comes up with another example of this misuse and abuse of correlational findings. Management consulting firm McKinsey did a correlational study of women and minorities on corporate boards and firm profitability. We know about the study because it found the two were positively correlated, the PC answer. McKinsey claimed to have shown having such board members led to more profits.

RealClearInvestigations has looked at these findings and identified a suspicion that, if anything, success led to hiring women and minority board members, rather than such board members leading to success. I am inclined to agree with their suspicion while freely admitting that the McKinsey data tell us only that the two occur together, nothing about causation. 

In the absence of better proof than McKinsey presents, I like Instapundit’s trope that firms which “Go woke, go broke.” An alternative explanation is that firms with ample profits often engage in non-essential activities that make their top managements “look good” in the eyes of those whose approval they seek. Economists refer to this as utilizing “organizational slack” for non-essential “feel good” activities.