As noted in the text, there may be several structural breaks in the U.S. economic time series dataset introduced in Section 21.1. Dummy variables are a good way of incorporating these shifts in the data.

a. Using dummy variables to designate three different periods related to the oil embargoes in 1973 and 1979, regress the log of personal consumption expenditures (LPCE) on the log of disposable personal income (LDPI). Has there been a change in the results? What is your decision about the unit root hypothesis now?

b. Several websites list the official economic cycles that may have affected the U.S. economic time series data discussed in Section 21.1. See, for example, http://www.nber.org/cycles/cyclesmain.html. Using the information here, create dummy variables indicating some of the major cycles and check the results of regressing LPCE on LDPI. Has there been a change?