Persuasion with Correlation Neglect: A Full Manipulation Result

We consider an information design problem in which a sender tries to persuade receiver that has “correlation neglect,” i.e., fails understand signals might be correlated. show with unlimited number of can fully manipulate the receiver. Specifically, induce hold any state-dependent posterior she wishes to. If only state-independent posterior, use correlated signals, but generally needs more involved correlation structures. (JEL D82, D83)