Can the Fed Control Real Interest Rates?

Three hypotheses concerning the controllability of rationally expected real interest rates are examined here. These hypotheses, which are suggested by recent literature, assert in different senses that the stochastic properties of expected real interest rates are independent of the Fed policy rule. We discuss the meaning and implications of the hypotheses, and how they might be tested. Evaluation of the hypotheses is attempted by examination of the Fed's "quasi-controlled experiments," historical changes in policy regimes, Granger-Sims causality tests, Barro unanticipated money regressions, and other methods. Questions as to the relevance of any such methods are discussed.