In the examples of autocorrelation in regression in this chapter, the Durbin-Watson statistic  was less than 2. What would it mean about the data if one found a significant value of ? Does this...

In the examples of autocorrelation in regression in this chapter, the Durbin-Watson statistic
 was less than 2. What would it mean about the data if one found a significant value of
? Does this explain why
 is typically less than 2 when the errors are dependent?

If the Durbin-Watson statistic is near 2 for the fit of an SRM to monthly data, have we proven that the errors are independent and meet the assumption of the SRM?