James Hamilton argues that oil price changes are still important for t
he macroeconomy, and that using a filter which screens out drops and m
easures increases relative to a reference level well represents the ef
fects of price changes when they are 'choppy', While this measure outp
erforms the oil price transformations considered in my paper, much of
its predictive power is coming from the pre-1986 (and pre-1973) part o
f the sample. With less than 10 years of choppy data, finding the best
oil price transformation will likely require analysis with cross-sect
ional components, Ruling out effects from oil price decreases is also
probably counterproductive to uncovering structure.