On multiple occasions on this blog, I have presented research that rationalizes some observation that at first sight appeared irrational, or at least bounded-rational. Maybe it is because of this obsession that economists have with rationality (and psychologists with irrationality).
Ran Spiegler claims these exercises can be very misleading. Such rationalization typically leans on three modifications of the standard model: some change in the information set, or in preferences, or there is a larger model where behavior is rational. Spiegler claims that this kind of rationalizing can be problematic. 1) It implies an actual change in the environment agents face. For some case studies at least, this is not appropriate. 2) It introduces new parameters, which need to be properly calibrated. 3) The rational equilibrium may have multiple equilibria. Which one should be selected? 4) The rationalized extension may not make sense. 5) Natural extensions may not make sense either. In other words, we should be really careful with rationalizations.