Infinite Variance and Research Strategy in Time Series Analysis

The possibilify that some time series fhof arise In economics have error or innovation ferms fhof come from infinite variance distribufions throws doubt on mosf of the classical methods of analysis. The evidence in favor of inhite-variance is discussed
and a variety of alfernotive explanations of fhe long-failed property of observed distribufions examined. Some of fhese alfernotive explanations are
based on mixfures of disfributions and suggesf dafa-transformations that reduce or remove the problem. Clipping of fhe series seems fo be a particularly useful technique and is applied to U.S. Treasury daily cash flow data.