The idea of rational expectations has revolutionized macroeconomics. Several authors believe that the idea can be easily tested by a simple econometric procedure. This paper, however, presents several reasons for questioning the relevance of such tests.