This paper estimates an interest rate reaction function for the US by using quarterly data spanning 1957:Q1-2010:Q3. We find evidence that US monetary policy can be described by the Taylor (1993) rule. Although Taylor''s original rule was specified under contemporaneous set-up, there is substantial evidence that forward looking specification can replace the original version without any loss of generality. Interest rate smoothing is another dimension that does not question the robustness of the rule. The only problem underlying in estimation process is the possibility of spurious regression that comes from the nonstationary behaviour of federal funds rate and inflation. We therefore suggest to examine the existence of Taylor rule as a cointegrating relationship between interest rate, inflation and output gap.