The purpose of this problem is to get you used to the concept of autocorrelation in a time series. You could do this with any time series, but here you should use the series of Walmart daily stock prices in the file P12_10.xlsx.
a. First, do it the quick way. Use the Autocorrelation procedure in StatTools to get a list of autocorrelations and a corresponding correlogram of the closing prices. You can choose the number of lags.
b. Now do it the more time-consuming way. Create columns of lagged versions of the Close variable—3 or 4 lags will suffice. Next, look at scatter plots of Close versus its first few lags. If the autocorrelations are large, you should see fairly tight scatters—that’s what autocorrelation is all about. Also, generate a correlation matrix to see the correlations between Close and its first few lags. These should be approximately the same as the autocorrelations from part a.
c. Create the first differences of Close in a new column. Now repeat parts a and b with the differences instead of the original closing prices—that is, examine the autocorrelations of the differences. They should be small, and the scatter plots of the differences versus lags of the differences should be shapeless swarms. This illustrates what happens when the differences of a time series variable have insignificant autocorrelations.
d. Write a short report of your findings.